Skip to main content

Full text of "Mind The World Order"

See other formats











Printed in the United States of America 

M. M. L. 


The conceptions presented in this book have 
grown out of investigations which began in the 
field of exact logic and its application to mathe- 
matics. The historic connection which exists be- 
tween mathematics and exact science on the one 
hand and conceptions of knowledge on the other, 
needs no emphasis : from Plato to the present day, 
all the major epistemological theories have been 
dominated by, or formulated in the light of, ac- 
companying conceptions of mathematics. Nor is 
the reason for this connection far to seek ; mathe- 
matics, of all human affairs, most clearly exhibits 
certitude and precision. If only one could come 
at the basis of this ideal character, the key-con- 
ceptions of epistemology might be disclosed- Thus 
every major discovery of theoretical mathematics, 
and every fundamental change in the manner in 
which this subject is conceived, is sure to find its 
sequel, sooner or later, in epistemology* Who- 
ever has followed the developments in logistic and 
mathematical theory in the last quarter-century 
can hardly fail to be convinced that the conse- 
quences of these must be revolutionary. It has 
been demonstrated, with a degree of precision and 
finality seldom attained, that the certitude of 
mathematics results from its purely analytic char- 
acter and its independence of any necessary con- 
nection with empirical fact. Its first premises are 



neither those self-evident truths of reason which 
inspired the continental rationalists to imitate the 
geometric method nor the principles of intuitive 
construction which, for Kant, assured a basis of 
application to all possible experience ; they are not 
even empirical generalizations, as Mill and other 
empiricists have thought. Rather, they are defini- 
tions and postulates which exhibit abstract con- 
cepts more or less arbitrarily chosen for the pur- 
poses of the system in question. Intrinsic connec- 
tion with experience is tenuous or lacking. 

Concurrently, developments in physical science, 
such as the theory of relativity, have emphasized 
the fact that, here too, abstractness and system- 
atic precision go together, and that exact deduc- 
tive procedures may give rise to no corresponding 
certainty about empirical nature. Logical integ- 
rity and concrete applicability are quite separate 
matters. The empirical truth of geometry even is 
not assured by its absolute validity as a deductive 
system, nor by any intuition of space, but de- 
pends upon further considerations. The nature of 
these is not wholly clear, but it is evident that ob- 
servation and the results of experiment must play 
some part in the determination of them. Thus the 
analytic character and abstractness of exact sys- 
tems, which assure to them that kind of certainty 
which they have, tend to divorce them from that 
empirical truth which is the object of natural 
science and the content of our possible knowledge 
of nature. 



We stand to-day so close to these developments 
that the far-reaching consequences of them may 
fail to impress us. It is not merely a change in one 
or two narrowly restricted disciplines at the far- 
thest remove from direct study of natural phe- 
nomena : whatever affects the basic subjects, such 
as mathematics and physics, is bound to be re- 
flected eventually throughout the whole of science* 
As a fact, this altered point of view is rapidly ex- 
tending to other branches, and the independence 
of the conceptual and the empirical is coming to 
be accepted as a commonplace. It is not too much 
to say, I think, that it becomes a matter of doubt 
whether the structure science builds is solidly 
based upon the earth, or is a mansion in some Pla- 
tonic heaven, or is only a kind of castle in the air. 
At least it appears that we must accept a kind of 
double-truth: there are the certainties, such as 
those of mathematics, which concern directly only 
what is abstract; and there are the presentations 
of our sense-experience to which we seek to apply 
them, but with a resultant empirical truth which 
may be no more than probable. The nature and 
validity of such empirical knowledge becomes the 
crucial issue. Traditional grounds of a priori 
truth have been, perforce, abandoned. What other 
grounds there may be; or whether without the a 
priori there can be any truth at all, must consti- 
tute our problem. 

So far as this is the case, the outstanding ques- 


tions concern the nature of our abstract concepts, 
such as those which figure in mathematics and 
theoretical physics, and the relation of them to 
concrete experience and to reality. Upon these 
points, the implications of current scientific de- 
velopments are nothing like so clear. If I could 
hope that I read these aright, and that something 
is here done toward rendering them explicit and 
consistent, I should, of course, be more than satis- 

The construction here attempted turns princi- 
pally upon three theses: (1) A priori truth is 
definitive in nature and rises exclusively from the 
analysis of concepts. That reality may be de- 
limited a priori, is due, not to forms of intuition or 
categories which confine the content of experience, 
but simply to the fact that whatever is denomi- 
nated "real" must be something discriminated in 
experience by criteria which are antecedently de- 
termined. () While the delineation of concepts 
is a priori, the application of any particular con- 
cept to particular given experience is hypotheti- 
cal; the choice of conceptual systems for such 
application is instrumental or pragmatic, and em- 
pirical truth is never more than probable. (3) 
That experience in general is such as to be capable 
of conceptual interpretation, requires no peculiar 
and metaphysical assumption about the conform- 
ity of experience to the mind or its categories ; it 
could not conceivably be otherwise. If this last 



statement is a tautology, then at least it must be 
true, and the assertion of a tautology is significant 
if it is supposed that it can be significantly denied. 
The development of these three theses will be 
found principally in Chapters III, VTII, and XI. 

Since this point of view will be likely to acquire 
some sort of label in any case, I shall venture to 
give it one myself and call it "conceptualistic 
pragmatism." Without the earlier conceptions of 
Peirce, James, and Dewey especially Peirce it 
would probably not have been developed. But these 
more orthodox pragmatists should not, of course, 
be made responsible for this view as a whole nor, 
particularly, for the doctrine of a priori truth 
which is included. 

In writing the book, I encountered a consider- 
able difficulty of exposition : with whatever one of 
its theses I should begin, the others would be more 
or less anticipated. In view of this difficulty, I 
have endeavored to keep the presentation as com- 
pact and swift as was compatible with clearness 
and proper emphasis. Controversial issues have 
been neglected except so far as discussion of them 
would contribute to the main development. And 
matters which lay to one side of the central theme 
but were still too important to be omitted alto- 
gether, have been covered briefly in appendices. 

To the graduate students in my seminary in 
the theory of knowledge in the last six years, my 
thanks are due for their critical discussion of these 


views, which have served both to bring out crucial 
points and to clarify my own conceptions* Even 
more important help of this sort came from dis- 
cussion with my colleagues of the Department of 
Philosophy in the University of California in the 
summer of 1926. To the officers of the Philosophi- 
cal Union there and of the University of Cali- 
fornia Press, I am indebted for permission to use 
again in these pages the materials which entered 
into the Howison Lecture for 1926, "The Prag- 
matic Element in Knowledge/' I am likewise in- 
debted to the editors of the Journal of Philosophy 
for permission to reprint brief excerpts from two 
articles, "The Structure of Logic and Its Rela- 
tion to Other Systems," and "A Pragmatic Con- 
ception of the A Priori. 5 * My friend and col- 
league, Professor E. G. Boring, has given me as- 
sistance of a kind most difficult to get by his 
critical appraisal of an earlier draft of the first 
four chapters and Appendix D. If these portions 
still leave much to be desired, at least they have 
been considerably improved as a result of his sug- 
gestions. Professor S. L. Quimby, of Columbia 
University, has assisted me with one of the illus- 
trations in Chapter VL As always, a principal 
stimulus and source of encouragement has been the 
association with my colleagues of the Harvard De- 

C/ I. L. 











TIONS 195 

















INDEX 441 




The general character of any philosophy is 
likely to be determined by its initial assumptions 
and its method. When Descartes proposed to 
sweep the boards clean by doubting everything 
which admitted of doubt and announced the initial 
criterion of certainty to be the inner light of hu- 
man reason, the distinguishing characteristics of 
the philosophic movement which resulted were 
thereby fixed* In similar fashion, the development 
from Locke to Hume is, for the most part, the 
logical consequence of the doctrine that the mind 
is a blank tablet on which experience writes. And 
when Kant proposed to inquire, not whether sci- 
ence is possible, but how it is possible, and identi- 
fied the possibility of science with the validity of 
synthetic judgments a priori, the successive at- 
tempts of the nineteenth century to deduce the 
major philosophic truths as presuppositions of 
experience was foreordained. 

Because method has this peculiar importance 
in philosophy, I believe that the reader of any 


philosophic book is entitled to know in advance 
what are the underlying convictions of this sort 
with which the writer sets out. It is right and 
proper that one should begin with some statement 
of program and method. 

It is I take it a distinguishing character of 
philosophy that it is everybody's business. The 
man who is his own lawyer or physician, will be 
poorly served; but everyone both can and must 
be his own philosopher. He must be, because phi- 
losophy deals with ends, not means. It includes 
the questions, What is good? What is right? What 
is valid? Since finally the responsibility for his 
own life must rest squarely upon the shoulder/* of 
each, no one can delegate the business of answer- 
ing such questions to another. Concerning the 
means whereby the valid ends of life may be at- 
tained, we seek expert advice. The natural sciences 
and the techniques to which they give rise, though 
they may serve some other interests also, are pri- 
marily directed to the discovery of such means. 
But the question of the ultimately valuable ends 
which shall be served, remains at once the most 
personal, and the most general of all questions. 

And everyone can be his own philosopher, be- 
cause in philosophy we investigate what we al- 
ready know. It is not the business of philosophy, 
as it is of the natural sciences, to add to the sum 
total of phenomena with which men are ac- 
quainted. Philosophy is concerned with what is 


already familiar. To know in the sense of fa- 
miliarity and to comprehend in clear ideas are, 
of course, quite different matters. Action precedes 
reflection and even precision of behavior com- 
monly outruns precision of thought fortunately 
for us. If it were not for this, naive common- 
sense and philosophy would coincide, and there 
would be no problem. Just this business of bring- 
ing to clear consciousness and expressing coher- 
ently the principles which are implicitly intended 
in our dealing with the familiar, is the distinctively 
philosophic enterprise. 

For instance, everybody knows the difference 
between right and wrong; if we had no moral 
sense, philosophy would not give us one. But who 
can state, with complete satisfaction to himself, 
the adequate and consistent grounds of moral 
judgment? Likewise, everyone knows the distinc- 
tion of cogent reasoning from fallacy. The study 
of logic appeals to no criterion not already pres- 
ent in the learner's mind. That logical error is, in 
the last analysis some sort of inadvertance, is an 
indispensable assumption of the study. Even if 
it should be in some part an unwarranted assump- 
tion, we could not escape it, for the very business 
of learning through reflection or discussion pre- 
sumes our logical sense as a trustworthy guide. 

That the knowledge sought in ethics and in 
logic is, thus, something already implicit in our 
commerce with the familiar, has usually been rec- 


ognized. But that the same is true in metaphysics, 
has not been equally clear. Metaphysics studies 
the nature of reality in general. Reality is pre- 
sumably independent of any principles of ours, 
in a sense in which the right and the valid may 
not be. At least initial presumption to the con- 
trary might be hopelessly prejudicial. Moreover, 
reality forever runs beyond the restricted field of 
familiar experience. What hope that cosmic rid- 
dles can be solved by self -interrogation ! The se- 
cret which we seek may be in some field which is 
not yet adequately explored or even opened to 
investigation. Or it may be forever beyond the 
reach of human senses. 

But it is not the business of philosophy to go 
adventuring beyond space and time. And so far 
as a true knowledge of the nature of reality de- 
pends on determining questions of phenomenal 
fact which are not yet settled, the philosopher has 
no special insight which enables him to pose as a 
prophet. We can do nothing but wait upon the 
progress of the special sciences. Or if speculate 
we must, at least such speculation is in no special 
sense the philosopher's affair. It is true that meta- 
physics has always been the dumping ground for 
problems which are only partly philosophic. 
Questions of the nature of life and mind, for ex- 
ample, are of this mixed sort. In part such issues 
wait upon further data from the sciences, from 
biology and physical-chemistry and psychology; 


in part they are truly philosophic, since they 
turn upon questions of the fundamental criteria 
of classification and principles of interpretation. 
No amassing of scientific data can determine these. 

If, for example, the extreme behaviorists in 
psychology deny the existence of consciousness on 
the ground that analysis of the "mental" must 
always be eventually in terms of bodily behavior, 
then it is the business of philosophy to correct 
their error, because it consists simply in a fallacy 
of logical analysis. The analysis of any immedi- 
ately presented X must always interpret this X 
in terms of its constant relations to other things 
to Y and Z. Such end-terms of analysis the 
Y and Z will not in general be temporal or spa- 
tial constituents of X 9 but may be anything which 
bears a constant correlation with it. It is as if one 
should deny the existence of colors because, for 
purposes of exact investigation, the colors must 
be defined as frequencies of vibratory motion. In 
general terms, if such analysis concludes by stat- 
ing "X is a certain kind of Y Z complex, hence 
X does not exist as a distinct reality ," the error 
lies in overlooking a general characteristic of 
logical analysis that it does not discover the 
"substance" or cosmic constituents of the phe- 
nomenon whose nature is analyzed but only the 
constant context of experience in which it will be 

So far, then, as the divergence of psychological 


theories, from behaviorism which interprets mind 
in terms of physical behavior to theories of the 
subconscious which assimilate much of physiolog- 
ical activity to mind, represents no dispute about 
experimental fact but only disparity of definition 
and methodological criteria, psychology and meta- 
physics have a common ground. The delineation 
of the fundamental concepts "mind" and "men- 
tal" is a truly philosophic enterprise. A similar 
thing might be discovered in the case of other 

Newly discovered scientific data might make 
such problems of fundamental concepts and classi- 
fication easier or more difficult but of itself it 
cannot solve them because, in the nature of the 
case, they are antecedent to the investigation. 
Such concepts are not simply dictated by the 
findings of the laboratory, or by any sort of sense- 
experience. Their origin is social and historical 
and represents some enduring human interest. It 
is the human mind itself which brings them to 
experience, though the mind does not invent them 
in a vacuum or cut them from whole cloth. The 
tendency to forget that initial concepts are never 
merely dictated by empirical findings is precisely 
what accounts for the absurd prejudice now hap- 
pily obsolescent that science is "just the report 
of facts." And this likewise helps to explain the 
common failure to distinguish between those cos- 
mological speculations which are not philosophic 


at all, because they are merely guesses at what 
future observation or experiment may reveal, 
from the legitimate and necessary philosophic 
question of a coherent set of fundamental cate- 
gories, such as "life" and "mind" and "matter," 
in terms of which experience may consistently and 
helpfully be interpreted. 

It would, of course, be captious to reserve this 
problem of initial concepts to philosophers, even 
though we should remember that, since everybody 
is to be his own philosopher, this merely means 
reserving them as general problems. The expert 
in the scientific field will have his special com- 
petence with respect to them; but they are not 
his exclusive property, because they are to be re- 
solved as much by criticism and reflection as by 
empirical investigation. Conversely, it would be 
pedantic if we should forbid the philosophic stu- 
dent to speculate concerning undetermined scien- 
tific fact. It is even questionable to deny the cap- 
tion "metaphysics" to those cosmological and 
ontological problems which have this partly specu- 
lative and partly critical or reflective character. 
Historically their title to the name is fairly good. 
All I wish to point out is that there is a real dis- 
tinction here between the speculative and the 
reflective elements; that this distinction coincides 
with a difference in the method by which resolu- 
tion of the problem is to be sought ; and that it is 
only the reflective element in such "metaphysical" 


problems which coincides in its nature and in the 
method of its solution with the problems of ethics 
and logic. 

With this explanation, I hope I shall cause no 
confusion if I say that it is only so far as they 
are thus critical and reflective that the problems 
of ontology and cosmology are truly philosophic ; 
and that metaphysics as a philosophic discipline 
is concerned with the nature of the real only so 
far as that problem is amenable to the reflective 
method and does not trench upon the field where 
only scientific investigation can achieve success. 
There are such reflective problems within any 
special science, and these may be said to consti- 
tute the philosophy of that science. There are also 
those problems of initial principle and criteria 
which are common to all the sciences and to the 
general business of life. These last are the prob- 
lems of philosophy proper. 

There is another sense in which metaphysics 
has often been speculative and departed from its 
proper philosophic business and method; that is, 
not by seeking to anticipate the science of the fu- 
ture, but through attempting by sheer force of ra- 
tional reflection to transcend experience altogether. 
Dogmatism is out of fashion since Kant.* But 

*Perhaps I should say, "has been out of fashion,** since just now 
we are being treated to various new forms of dogmatism* But this, 
I take it, comes partly as a counsel of despair, and partly it represents 
a reaction against the often exaggerated claims of "idealism" and 
post-Kantian "criticism" to be able to proceed a priori without refer- 
ence to particular results of the empirical sciences. 


that philosophic legerdemain which, with only ex- 
perience for its datum, would condemn this expe- 
rience to the status of appearance and disclose 
a reality more edifying, is still with us. The mo- 
tives of this attitude are, indeed, ingrained in 
human nature, and I am reluctant to lay hands 
on that idealism which has played the role of 
Father Parmenides to all the present generation 
of philosophers. But at least we must observe that 
such metaphysics turns away from one type of 
problem which is real and soluble to another which 
may not be. Even if all experience be appearance, 
and all every-day thought and truth infected with 
contradiction, at least it must be admitted that 
some appearances are better than others. The 
mundane distinction of real and unreal within ex- 
perience has its importance and calls for formu- 
lation of its criteria. It may be that Reality, with, 
a capital R, the concrete-universal Reality which 
transcends all particular phenomena and under- 
lies them, is a kind of philosophical ignis fatuus. 
Perhaps the idea of "whole" applies only within 
experience, and no whole can validly be conceived 
except such as stands in contrast with something 
else and has concrete bounds. Perhaps the whole 
of Reality is, as Kant thought, an inevitable idea 
but also a necessarily empty one, to remind us 
forever of the more which is to be learned and 
connected with our previous knowledge. But 
whether this be so or not, there is the less ambitious 


and more important problem of determining the 
criteria by which the adjective "real" is correctly 
applied the problem of the abstract universal. 
And if any be inclined to think that this question 
is too simple or too meager for a philosophic dis- 
cipline, I shall hope to indicate his error. 

A metaphysics which takes this as its problem 
will remain strictly within the reflective method. 
It will seek to determine the nature of the real, as 
ethics seeks to determine the good, and logic, the 
valid, purely by critical consideration of what 
does not transcend ordinary experience. That is, 
it will seek to define "reality," not to triangulate 
the universe. It will be concerned with the formu^ 
lation of principles, but of principles already im- 
manent in intelligent practice. A person with no 
sense of reality (other-worldly philosophers, for 
example) will not acquire one by the study of 
metaphysics. And by no possibility can such in- 
vestigation reveal reality as something esoteric 
and edifying and transcendent of ordinary ex- 
perience. Any metaphysics which portrays reality 
as something strangely unfamiliar or beyond the 
ordinary grasp, stamps itself as thaumaturgy, 
and is false upon the face of it. 

The problem of 'a correctly conceived meta- 
physics, like the problem of ethics and of logic, is 
one to be resolved by attaining to clear and co- 
gent self -consciousness. As it turns out, the prob- 
lem of metaphysics is "the problem of the cate- 


gories,"* The reason for this lies in a curious 
complexity of the meaning of "reality. 9 * Logical 
validity is at bottom of one single type. And per- 
haps the good and the right are relatively simple 
in their ultimate nature. But the adjective "real" 
is systematically ambiguous and can have a sin- 
gle meaning only in a special sense. The ascrip- 
tion of reality to the content of any particular ex- 
perience is always eliptical: some qualification 
material reality, psychic reality, mathematical 
reality is always understood. And whatever is 
real in one such sense will be unreal in others. 
Conversely, every given content of experience is 
a reality of some sort or other ; so that the prob- 
lem of distinguishing real from unreal, the prin- 
ciples of which metaphysics seeks to formulate, is 
always a problem of right understanding, of re- 
ferring the given experience to its proper cate- 
gory. The mirage, for example, though not real 
trees and water, is a real state of atmosphere 
and light ; to relegate it to the limbo of nothing- 
ness would be to obliterate a genuine item of the 
objective world. A dream is illusory because the 
dreamer takes its images for physical things; 
but to the psychologist, interested in the scien- 
tific study of the mental, just these experienced 
images, occurring in just this context of other cir- 
cumstance, constitute a reality to be embraced 

*A more logical terminology would qualify this as the "categories 
of reality/' and would distinguish these from the "categories of 
of value.** 


under law and having its own indisputable place 
in the realm of fact. The content of every experi- 
ence is real when it is correctly understood, and 
is that kind of reality which it is then interpreted 
to be. Metaphysics is concerned to reveal just 
that set of major classifications of phenomena, 
and just those precise criteria of valid understand- 
ing, by which the whole array of given experience 
may be set in order and each item (ideally) as- 
signed its intelligible and unambiguous place. 

So understood, the principles of the categories, 
which metaphysics seeks, stand, on the one side, 
in close relation to experience and can not mean- 
ingfully transcend it. But on the other side or 
in a different sense they stand above or before 
experience, and are definitive or prescriptive, and 
hence a priori. 

Whatever principles apply to experience must 
be phrased in terms of experience. The clues to 
the categorial* interpretation the correct under- 
standing of any presentation of sense must be 
empirical clues. If they are not contained within 
that segment of experience which constitutes the 
phenomenon itself, then they must be discover- 
able in its relation to other empirical fact. If 
the dream or illusion is not betrayed by inter- 
nal evidence, then its true nature must be dis- 
closed by the conjunction with what precedes or 

* M Categorial " is used throughout with the meaning ** pertaining 
to the categories." This avoids possible confusion wiih "categorical," 
meaning specifically "unconditional, not hypothetical." 


follows. But while the distinguishing marks of 
reality of any particular sort are thus experi- 
mental, the principles by which the interpretation 
or classification is made are prior to the experi- 
ence in question. It is only because the mind is 
prepared to judge it real or unreal according as 
it bears or fails to bear certain marks, that inter- 
pretation of the given is possible at all, and that 
experience can be understood. 

It is through reflective examination of expe- 
rience (more particularly of our own part in it or 
attitude toward it) that we may correctly formu- 
late these principles of the categories, since they 
are implicit in our practical dealings with the em- 
pirically given. But they are not empirical gen- 
eralizations in the sense that some later experience 
may prove an exception and thus invalidate them. 
They formulate an attitude of interpretation or 
discrimination by which what would be excep- 
tional is at once thrown out of court. For exam- 
ple, no experience of the physical can fail to bear 
those marks the absence of which would bar the 
given content of experience from interpretation 
as physical reality. The formulation of our de- 
liberately taken, and consistently adhered to, at- 
titude of interpretation constitutes a categorial 
definition of "the physical/ 5 Such a categorial 
definitive principle forbids nothing in the way of 
experience ; it prohibits neither illusion nor sense- 
less dream. Thus such principles are not mate- 


rial truths : they can be a priori knowable with 
certainty in advance of experience precisely be- 
cause they impose no limitation upon the given 
but, as principles of interpretation, nevertheless 
condition it as a constituent of reality. It will be 
the thesis of a later chapter that the a priori has 
in general this same character of definitive prin- 
ciple, which does not limit the content of the given, 
We shall find in this nature of the a priori the 
solution of many traditional problems of the the- 
ory of knowledge. 

So conceived, the principles which formulate 
criteria of the real, in its various types, are a 
priori in precisely the same sense as are the can- 
ons of ethics and of logic. Experience does not 
itself determine what is good or bad, or the na- 
ture of goodness, nor does it determine what is 
valid or invalid, or the nature of logical validity. 
Equally it does not determine what is real or un- 
real (in any particular sense), or the nature "of 
reality. Experience does not categorize itself. The 
criteria of interpretation are of the mind; they 
are imposed upon the given by our active attitude. 

The main business of a sound metaphysics is, 
thus, with the problem of the categories; the 
formulation of the criteria of reality, in its vari- 
ous types. It is to the shame of philosophy that 
these problems, which by their nature must be ca- 
pable of precise solution since they require only 
persistent regard for fact and self-conscious ex- 


amination of our own grounds of judgment, have 
been so generally neglected. Just this common dis- 
regard of verifiable fact and mundane criteria of 
the real is largely responsible for that quagmire 
of incertitude and welter of the irrelevant and 
vague which at present bears the name of meta- 
physics. The problems of the categories admit of 
as much real progress as those of logic; in fact, 
they are problems of the same general type- We 
may congratulate ourselves, I think, that a grow- 
ing interest in such study, in this reflective or 
phenomenalistic or critical spirit, is one of the 
characteristic of the present period in philoso- 

The definition of the real in general, and the 
picturing of reality as a whole, are subordinate 
matters ; and perhaps, as has been suggested, the 
second of these is not possible. The word "real" 
has a single meaning, of course, in the same sense 
that "useful 5 ' or any other such elliptical term has 
a single meaning. Nothing is useful for every pur- 
pose, and perhaps everything is useful for some 
purpose. A definition of "usef ul" in general would 
not divide things into two classes, the useful and 
the useless. Nor could we arrive at such a defini- 
tion by attempting to collect all useful things into 
a class and remark their common characters, since 
we should probably have everything in the class 

*I have in mind, as examples, Whitehead*s ** Concept of Nature" 
and "Principles of Natural Knowledge," Russell's "Analysis of 
Mind," and Broad's "Scientific Thought/' 


and nothing outside it to represent the useless. 
Instead, we should first have to consider the dif- 
ferent types of usefulness or of useful things and 
then discover, if possible, what it is that charac- 
terizes the useful as contrasted with the useless in 
all these different cases. We should find, of course, 
that it was not some sense-quality but a relation 
to an end which was the universal mark of useful- 
ness. Similarly, to arrive at a general definition 
of "the real" it would not do to lump together all 
sorts of realities in one class and seek directly for 
their common character* Everything in this class 
would be at once real, in some category, and un- 
real in others. And nothing would be left outside 
it. The subject of our generalization must be, in- 
stead, the distinction real-unreal in all the differ- 
ent categories. What definition of reality in gen- 
eral we might thus arrive at, we need not pause to 
inquire. Obviously it would be found to embrace 
some relation to empirical givenness in general or 
to our interpreting attitude, or to something in- 
volving both of these, rather than any particular 
and distinguishing empirical characteristics. 

That in any case a successful definition of the 
real in general would not carry us far in any cos- 
mological attempt to plumb the deeps of the uni- 
verse, is evident from the fact that it would de- 
limit reality in intension only, and would leave 
quite undetermined the particular content of real- 
ity in extenso. The total picture of reality can be 


drawn only when the last experience of the last 
man, and the final facts of science, are summed 
up. Why cosmology in this sense should be sup- 
posed to be the business of the philosopher or of 
anyone else I cannot see. In the nature of the 
case, it must be a cooperative enterprise, and pre- 
sumably one that is always incomplete. 

What we have here seen concerning the sig- 
nificance of "the real" will have its importance 
for certain topics discussed in later chapters* But 
our immediate interest in it lies in the fact that it 
brings metaphysics which threatened to prove 
an exception back into line with other branches 
of philosophy with respect to the method by which 
it should be pursued. It is only in and through the 
general course of human experience that we have 
a content for our philosophic thinking, and the 
significance of philosophic truth lies always in its 
application to experience. But it is experience 
from a certain point of view, or a certain aspect 
of it, with which we are concerned. Ethics cannot 
tell us how much of life is good, what particular 
sins are committed, or what proportion of men 
are moral; nor does metaphysics describe the 
course of the universe or determine the extent and 
the particulars of the real. It is the logical es- 
sence of goodness, the canons of validity, the cri- 
teria of the beautiful, and likewise, the principles 
of the distinction of real from unreal, that phi- 
losophy may hope to formulate. These criteria 


and principles, the mind itself brings to experi- 
ence in its interpretation, its discriminations, and 
its evaluation of what is given. Thus philosophy 
is, so to speak, the mind's own study of itself in 
action; and the method of it is simply reflective* 
It seeks to formulate explicitly what from the be- 
ginning is our own creation and possession. 

However, I should not like to appear to defend 
the notion that such analysis is a simple matter or 
that it requires only to express in precise terms 
the principles of common-sense. As has often 
enough been emphasized, common-sense is itself a 
naive metaphysics and one which frequently 
breaks down on examination. Just as naive moral- 
ity may become confused before the dialectical at- 
tack, so common-sense categories of reality fail in 
crucial cases to meet the tests of consistency and 
accord with intelligent practice. It is true in meta- 
physics, as it is in ethics and in logic, that while 
valid principles must be supposed somehow im- 
plicit in the ordinary intercourse of mind with 
reality, they are not present in the sense of being 
fatally adhered to. If they were, the philosophic 
enterprise would have no practical value. Self- 
consciousness may be an end in itself, but if it 
did not have eventual influence upon human ac- 
tion it would be a luxury which humanity could 
not afford. That we coincide in our logical sense, 
does not make logic a work of supererogation. 
No more does coincidence in our ultimate sense of 


reality and in our categories render metaphysical 
discussion nugatory. Just as the study of logic 
may conduce to cogency of thought, and ethics 
contribute to greater clarity and consistency in 
moral judgment, so too the elucidation of meta- 
physical problems may contribute to the precision 
and adequacy of our interpretation of the real ; it 
may even serve, on occasion, to work improvement 
in the concepts of the special sciences. Philosophy 
cannot be merely a verbally more precise render- 
ing of common-sense, nor a direct generalization 
from actual practice. Though it rises from what 
is implicit in experience, its procedure must be 
critical, not descriptive. So far as it is to be of 
use, it must assume the function of sharpening 
and correcting an interpretation which has al- 
ready entered into the fabric of that experience 
which is its datum. Logical principles aim to re- 
place the uncritical moral sense, ethics, our naive 
morality, and metaphysics, our unreflective on- 
tological judgments. Such an enterprise is no sim- 
ple matter of formulating the obvious. 

The reflective method must, or course, be dia- 
lectical in the Socratic-Platonic, not the Hegel- 
ian, sense. It accords with the Socratic presump- 
tion that the truth which is sought is already im- 
plicit in the mind which seeks it, and needs only 
to be elicited and brought to clear expression. It 
accords, further, in the recognition that it is defi- 
nitions or "essences" which are the philosophic 


goal. And it likewise recognizes that the hope of 
agreement between minds, to be reached by philo- 
sophic discussion, must rest upon the presump- 
tion that this accord somehow exists already. 

Historically, however, the dialectical method 
has been overlaid with all sorts of addenda, and 
perverted by extraneous assumptions which are 
fallacious. So that I should choose the name "re- 
flective" as less liably to unwarranted interpreta- 
tion. It does not follow from the dialectical method 
that the basis of the accord between minds repre- 
sents some universal pattern of human reason, 
apart from the world of sense in which we live; 
nor that the mind has access to some realm of 
transcendent concepts which it recovers, of its 
own powers, at the instigation of experience ; nor 
that agreement of minds presumes initial princi- 
ples which are self-evident. It does not even follow 
that the agreement which we seek is already im- 
plicity complete in all respects. To all such no- 
tions there is an alternative, to account for this 
agreement between minds, which is simple and 
even obvious. The coincidence of our fundamental 
criteria and principles is the combined result of 
the similarity of human animals, and of their pri- 
mal interests, and the similarities of the experience 
with which they have to deal. More explicitly, it 
represents one result of the interplay between 
these two ; the coincidence of human modes of be- 
havior, particularly when the interests which such 
behavior serves involve cooperation. 


Our categories are guides to action. Those atti- 
tudes which survive the test of practice will reflect 
not only the nature of the active creature but the 
general character of the experience he confronts. 
Thus, indirectly, even what is a priori may not be 
an exclusive product of "reason," or made in 
Plato's heaven in utter independence of the world 
we live in. Moreover, the fact that man survives 
and prospers by his social habits serves to ac- 
centuate and perfect agreement in our basic at- 
titudes. Our common understanding and our com- 
mon world may be, in part, created in response to 
our need to act together and to comprehend one 
another. Critical discussion is but a prolongation 
of that effort which we make to extend the bounds 
of successful human cooperation. It is no more 
necessary to suppose that agreement in funda- 
mental principles is completely ready-made than 
it is to suppose that infants must already have 
precisely those ideas which later they find words 
to express. Indeed our categories are almost as 
much a social product as is language, and in 
something like the same sense. It is only the pos- 
sibility of agreement which must be antecedently 
presumed. The "human mind" is a coincidence of 
individual minds which partly, no doubt, must be 
native, but partly is itself created by the social 
process. Even that likeness which is native would 
seem to consist in capacities and tendencies to ac- 
tion, not in mental content or explicit modes of 
thought. That the categories are fundamental in 


such wise that the social process can neither cre- 
ate nor alter them, is a rationalistic prejudice 
without foundation. There is much which is pro- 
found and true in traditional conceptions of the 
a priori. But equally it should be clear that there 
is much in such conceptions which smacks of magic 
and superstitious nonsense. Particularly it is im- 
plausible that what is a priori can be rooted in a 
"rational nature of man" which is sometimes mi- 
raculous and beyond the bounds of psychological 
analysis and genetic explanation- 
It may be pointed out also that if we recognize 
critical reflection or dialectic as the only method 
which holds promise in philosophy, we do not 
thereby commit ourselves to the assumption that 
coherence or internal consistency is the only test, 
or a sufficient test, of philosophic truth* In phi- 
losophy, as elsewhere, consistency is only a nega- 
tive test of truth : it is possible, however unlikely, 
to be consistently in error. Consistency would be 
a sufficient test only if we should suppose that 
there is nothing external to our logic which we 
must be true to. The reflective method does not 
take it for granted that all fact follows, Hegelian- 
fashion, from the logical structure of thought 
itself. As has been suggested, it does not even pre- 
suppose that what is a priori and of the mind 
our categorial attitude of interpretation is com- 
pletely independent of the general character of 


It is of the essence of the dialectical or reflec- 
tive method that we should recognize that proof, 
in philosophy, can be nothing more at bottom 
than persuasion. It makes no difference what the 
manner of presentation should be, whether de- 
ductive from initial assumptions, or inductive 
from example, or merely following the order dic- 
tated by clarity of exposition. If it be deductive, 
then the initial assumptions cannot coerce the 
mind. There are no propositions which are self- 
evident in isolation. So far as the deductive pres- 
entation hopes to convince of what was not pre- 
viously believed, it must either seek out initial 
agreements from which it may proceed, or as is 
more frequently the case the deductively first 
propositions must be rendered significant and ac- 
ceptable by exhibiting the cogency and general 
consonance with experience of their consequences. 
If the method be inductive from example, then 
the principles to be proved are implicit in the as- 
sumption that cited examples are veridical and 
typical and genuinely fall under the category to 
be investigated. There can be no Archimedean 
point for the philosopher. Proof, he can offer only 
in the sense of so connecting his theses as to ex- 
hibit their mutual support, and only through ap- 
peal to other minds to reflect upon their experi- 
ence and their own attitudes and perceive that he 
correctly portrays them. If there be those minds 
which find no alternatives save certainty, apart 


from all appeal to prior f act, or skepticism, then 
to skepticism they are self-condemned. And much 
good may it do them! As philosophers, we have 
something we must be faithful to, even if that 
something be ourselves. If we are perverse, it is 
possible that our philosophy will consist of lies. 

Already this introductory analysis of method is 
too long. But the conception of the a priori here 
suggested is a novel one: a little further discus- 
sion may have its value by way of anticipating 
briefly what is to follow. 

If Philosophy is the study of the a priori, and 
is thus the mind's formulation of its own active 
attitudes, still the attitude which is the object of 
such study is one taken toward the content of an 
experience in some sense independent of and 
bound to be reflected in the attitude itself. What 
is a priori it will be maintained is prior to ex- 
perience in almost the same sense that purpose is. 
Purposes are not dictated by the content of the 
given; they are our own. Yet purposes must take 
their shape and have their realization in terms of 
experience ; the content of the given is not irrele- 
vant to them. And purposes which can find no ap- 
plication will disappear. In somewhat the same 
fashion what is a priori and of the mind is prior 
to the content of the given, yet in another sense 
not altogether independent of experience in gen- 

It is an error common to rationalism and to 


pure empiricism that both attempt an impossible 
separation of something called the mind from 
something else called experience. Likewise both 
treat of knowledge as if it were a relation of the 
individual mind to external object in such wise 
that the existence of other minds is irrelevant; 
they do not sufficiently recognize the sense in 
which our truth is social. Traditional rational- 
ism,* observing that any principles which should 
serve as ultimate criterion or determine cate- 
gorial interpretation must be prior to and inde- 
pendent of the experience to which it applies, has 
supposed that such principles must be innate and 
so discoverable by some sort of direct inspection. 
If a canon of their truth is requisite, this must be 
supplied by something of a higher order than ex- 
perience, such as self -evidence or the natural light 
of reason. The mistakes of this point of view are 
two. In the first place, it assumes that mind is im- 
mediate to itself in a sense in which the object of 
experience is not. But what other means have we 
of discovering the mind save that same experi- 
ence in which also external objects are presented? 
And if the object transcends the experience of it, 
is not this equally true of the mind? The single 
experience exhausts the reality of neither. Any 
particular experience is a whole within which that 
part or aspect which represents the legislative or 

The rationalism (if that term is justified) of post-Kantian ideal- 
ism rests upon different assumptions and proceeds by different 
methods. It is not here in point. 


categorial activity of mind and that which is 
given content, independent of the mind's inter- 
pretation, are separable only by analysis. We have 
no higher faculty or more esoteric experience 
through which the mind discovers itself. And sec- 
ond, rationalism fallaciously assumes that what is 
prior to, or legislative for, the particular experi- 
ence must be likewise independent of experience 
in general. Though categorial principle must, in 
the nature of the case, be prior to the particular 
experience, it nevertheless represents an attitude 
which the mind has taken in the light of past ex- 
perience as a whole, and one which would even be 
susceptible of change if confronted with some 
pervasive alteration in the general character of 
what is presented. An example here may be of 
service: It is an a priori principle that physical 
things must have mass. By this criterion, they are 
distinguished from mirror-images and illusion. 
Since this is so, no particular experience could 
upset this principle, because any experience in 
which it should be violated would be repudiated 
as non-veridical or "not correctly understood." 
That is, by the principle itself, the phenomenon 
must be referred to some other category than the 
physical. In that sense, the truth of the princi- 
ple is independent of the particular phenomenon. 
But a world in which we should experience phe- 
nomena having a persistence and independence 
not characteristic of imagination, and a coherence 


not characteristic of our dreams, but things which 
would still not be amenable to any gravitational 
generalizations, is entirely conceivable. In such a 
world our a priori principle would not be ren- 
dered false since it is definitive of the physical; 
but the category "physical" might well be useless. 
(Incidentally it may be pointed out that this cri- 
terion of the physical is a historical and social 
product. Aristotle and the ancients knew it not.) 

Though we bring the a priori principle, as cri- 
terion, to any particular experience, yet this legis- 
lative attitude of mind is clearly one which is 
taken because, our experience on the whole being 
what it is, this principle helps to render it intel- 
ligible, and behavior in accord with it is normally 
successful. The mind must bring to experience 
whatever serve as the criteria of interpretation 
of the real, as of the right, the beautiful, and the 
valid. The content of experience cannot evaluate 
or interpret itself. Nevertheless the validity of 
such interpretation must reflect the character of 
experience in general, and meet the pragmatic test 
of value as a guide to action. 

The fallacy of pure empiricism is the converse 
of that which rationalism commits. In seeking to 
identify the real with what is given in experience, 
apart from construction or interpretation by the 
mind, and to elicit general principles directly 
from the content of experience, empiricism con- 
demns itself to a vicious circle. Experience as it 


comes to us contains not only the real but all the 
content of illusion, dream, hallucination, and mis- 
apprehension. When the empiricist supposes that 
laws or principles can be derived simply by gen- 
eralization from experience, he means to refer 
only to veridical experience, forgetting that with- 
out the criterion of legislative principle experi- 
ence cannot first be sorted into veridical and 

It is this vicious circle which makes inevitable 
the historical denouement of empiricism in Hume's 
skepticism. Berkeley pointed out that the real 
cannot be distinguished from the unreal by any 
relation between the idea in the mind and an in- 
dependent object, but only by some relation within 
experience itself. In this, of course, he is right, 
whether we agree with his idealism or not: mind 
cannot transcend itself and discover a relation of 
what is in experience to what is not. Berkeley 
then seeks to indicate our actual empirical cri- 
teria: the real in experience is distinguished (1) 
by that independence of the will which is exhib- 
ited in the content of perception as contrasted 
with imagination, (2) by the greater liveliness of 
perception, (3) by the interconnection of verid- 
ical perceptions according to the "laws of na- 
ture." Obviously only the last of these is sufficient 
in critical cases such as hallucination and errors 
of observation. Hume wrecks the empiricist struc- 
ture when he points out that such "laws of na- 


ture" cannot be derived by generalization from 
experience. For this, the distinction of necessary 
from contingent would be requisite. The basis of 
this distinction is not to be found in the content 
of experience; it is of the mind* Generalization 
from experience always presumes that the cate- 
gorial interpretation already has been made. 
Laws which characterize all experience, of real 
and unreal both, are non-existent, and would in 
any case be worthless. 

It is obvious that similar considerations hold 
for the other problems of philosophy. The nature 
of the good can be learned from experience only 
if the content of experience be first classified into 
good and bad, or grades of better and worse. 
Such classification or grading already involves 
the legislative application of the same principle 
which is sought. In logic, principles can be elicited 
by generalization from examples only if cases of 
valid reasoning have first been segregated by some 
criterion. It is this criterion which the generaliza- 
tion is required to disclose. In esthetics, the laws 
of the beautiful may be derived from experience 
only if the criteria of beauty have first been cor- 
rectly applied. 

The world of experience is not given in experi- 
ence: it is constructed by thought from the data 
of sense. This reality which everybody knows re- 
flects the structure of human intelligence as much 
as it does the nature of the independently given 


sensory content. It is a whole in which mind and 
what is given to mind already meet and are inter- 
woven* The datum of our philosophic study is not 
the "buzzing, blooming confusion" on which the 
infant first opens his eyes, not the thin experience 
of immediate sensation, but the thick experience 
of every-day life. 

This experience of reality exists only because 
the mind of man takes attitudes and makes inter- 
pretations. The buzzing, blooming confusion 
could not become reality for an oyster. A purely 
passive consciousness, if such can be conceived, 
would find no use for the concept of reality, be- 
cause it would find none for the idea of the ^?ireal ; 
because it would take no attitude that could be 
balked, and make no interpretation which conceiv- 
ably could be mistaken. 

On the other hand, we can discover mind and 
its principles only by analysis in this experience 
which we have. We cannot, unless dogmatically, 
construct experience from a hypothetical and 
transcendent mind working upon a material which 
likewise is something beyond experience. We can 
only discover mind and what is independently 
given to it by an analysis within experience itself. 
And it is only because mind has entered into the 
structure of the real world which we know and the 
experience of everyday, that analysis, or any at- 
tempted knowledge, may discover it. 

In finding thus that the principles and criteria 


which philosophy seeks to formulate must be sig- 
nificant at once of experience and of our active 
attitudes, the reflective method inevitably is prag- 
matic also. Concepts and principles reveal them- 
selves as instruments of interpretation; their 
meaning lies in the empirical consequences of the 
active attitude* The categories are ways of deal- 
ing with what is given to the mind, and if they had 
no practical consequences, the mind would never 
use them. Since philosophy seeks to formulate 
what is implicit in mind's every-day interpreta- 
tions, we may test the significance of any philo- 
sophic principle, and pave the way for determin- 
ing its truth, if we ask : How would experience be 
different if this should be correct than if it should 
be false? or, How differently should we orient our- 
selves to experience and deal with it if this should 
be so than if it should be not so? 

Metaphysical issues which supposedly concern 
what is transcendent of experience altogether, 
must inevitably turn out to be issues wrongly 
taken. For example, if one say as Mr* Broad has 
recently said* that scientific reality of perdur- 
ing electrons or what not, is something which at 
best is probable only, since it does not enter our 
direct experience of "sensa," then I think we may 
justly challenge him as Berkeley challenged 
Locke: Why not a world of sensa with nothmg 
behind them? What makes "scientific reality" even 
*" Scientific Thought," see esp. pp. 268 ff. 


probable if direct experience could be the same 
without its existence? Unless the modern physicist 
hopelessly deludes himself, does not the existence 
of electrons mean something verifiable in the 
laboratory? Otherwise, would he not be con- 
strained to answer any question about electrons as 
Laplace is reputed to have answered Napoleon's 
question about G-od that he had no need of this 
hypothesis? But if the existence or non-existence 
of "scientific reality" makes certain verifiable dif- 
ferences in experience, then these empirical cri- 
teria are the marks of the kind of reality which 
can be predicated of it. They are the "cash-value" 
of the category ; they constitute what it means to 
be real in just the way that electrons can be real. 
"Scientific reality" is either an interpretation of 
certain parts and aspects of experience or it is a 
noise, signifying nothing. 

The totality of the possible experiences in 
which any interpretation would be verified the 
completest possible empirical verification which is 
conceivable constitutes the entire meaning which 
that interpretation has. A predication of reality 
to what transcends experience completely and in 
every sense, is not problematic ; it is nonsense. 

Perhaps another illustration may make the 
point more clear. Occasionally philosophers amuse 
themselves by suggesting that the existences of 
things are intermittent ; that they go out when we 
cease to notice them and come into being again at 


the moment of rediscovery. The answer is not 
given by any question-begging reference to the 
independent object or to the conservation of mat- 
ter. What we need to inquire is why this notion of 
permanent objects was ever invented. If nothing 
in experience would be different whether the ex- 
istence of things should be intermittent or con- 
tinuous, what character of experience is predi- 
cated by their "permanence 93 ? When we have an- 
swered to such questions, we have discovered the 
whole meaning of "permanent existence 59 and 
nothing further, unless paradox of language, re- 
mains to be discussed. Reflection upon experience 
and our attitude to what is given cannot discover 
what is not implicitly already there and there is 
nothing else which philosophic reflection can hope 
to disclose. 

To sum up, then: The reflective method is em- 
pirical and analytic in that it recognizes experi- 
ence in general as the datum of philosophy. But 
it is not empirical in the sense of taking this ex- 
perience to coincide with data of sense which are 
merely given to the mind. Nor is it analytic in the 
sense of supposing that experience is complete and 

Rather, it finds that philosophy is particularly 
concerned with that part or aspect of experience 
which the mind contributes by its attitude of in- 
terpretation. In thus recognizing that the prin- 
ciples which are sought are in some sense a priori, 
it is rationalistic. 


It is not rationalistic, however, in the sense of 
presuming the mind as a Procrustean bed into 
which experience is forced, or as an initial datum 
which can be assumed or its findings known apart 
from sense-experience. Nor does it presume the 
"rational human mind" as something completely 
identical in and native to all human beings, or as 
a transcendent entity which, even if it lived in 
some other world of sense, would still possess pre- 
cisely the same categories and pattern of intelli- 

The reflective method is pragmatic in the same 
sense that it is empirical and analytic. It supposes 
that the categories and principles which it seeks 
must already be implicit in human experience and 
human attitude. The significance of such funda- 
mental conceptions must always be practical be- 
cause thought and action are continuous, and be- 
cause no other origin of them can be plausible 
than an origin which reflects their bearing on ex- 
perience. Further, it claims for philosophy itself 
the pragmatic sanction that reflection is but a 
further stretch of that critical examination of our 
own constructions and interpretations by which 
we free them from inconsistency and render them 
more useful. Since experience is not just given 
but is in part a product of the mind, philosophy 
itself may work some alteration of the active atti- 
tude by which the given in experience is met and 
moulded. But the reflective method is not a or need 


not be, pragmatic in the sense of supposing, as 
current pragmatism sometimes seems to do, that 
the categories of biology and psychophysics have 
some peculiar advantage for the interpretation of 
the practical attitudes of thought. 

The reflective method necessarily leads to the 
repudiation of any reality supposed to be trans- 
cendent of experience altogether. A true philo- 
sophic interpretation must always follow the clues 
of the practical reasons for our predications. A 
philosophy which relegates any object of human 
thought to the transcendent, is false to the human 
interests which have created that thought, and to 
the experience which gives it meaning. Philosophic 
truth, like knowledge in general, is about experi- 
ence, and not about something strangely beyond 
the ken of man, open only to the seer and the 
prophet. We all know the nature of life and of the 
real, though only with exquisite care can we tell 
the truth about them. 



The presumption from which we set out is that 
it is the business of philosophy to analyze and in- 
terpret our common experience, and by reflection, 
to bring to clear and cogent expression those prin- 
ciples which are implicit because they are brought 
to experience by the mind itself. Philosophy is the 
study of the a priori. It seeks to reveal those cate- 
gorial criteria which the mind applies to what is 
given to it, and by correct delineation of those 
criteria to define the good, the right, the valid, 
and the real. 

The attempt, however, so to approach the prob- 
lems of philosophy leads at once to outstanding 
questions concerning the nature of knowledge, 
solution of which seems prerequisite. The distinc- 
tion between what is a priori and what is not, is 
here presumed; as is also the correlative distinc- 
tion between mind, or what mind brings to experi- 
ence, and some other element, presumably inde- 
pendent of the mind's activity and responsible for 
other parts or aspects of experience. Have we a 
right to these distinctions? What are the grounds 
on which they can be drawn? How, in these terms, 



are knowledge and experience constituted? It is 
with such questions that the remainder of this book 
is concerned. 

Its principal theses are the following: (1) The 
two elements to be distinguished in knowledge are 
the concept, which is the product of the activity 
of thought, and the sensuously given, which is in- 
dependent of such activity. (2) The concept gives 
rise to the a priori ; all a priori truth is definitive, 
or explicative of concepts. (3) The pure concept 
and the content of the given are mutually inde- 
pendent; neither limits the other. (4) Empirical 
truth, or knowledge of the objective, arises 
through conceptual interpretation of the given. 
(5) The empirical object, denoted by a concept, 
is never a momentarily given as such, but is some 
temporally-extended pattern of actual and pos- 
sible experience. (6) Hence the assignment of any 
concept to the momentarily given (which is char- 
acteristic of perceptual knowledge) is essentially 
predictive and only partially verified. There is no 
knowledge merely by direct awareness. (7) Actual 
experience can never be exhaustive of that pat- 
tern, projected in the interpretation of the given, 
which constitutes the real object. Hence all em- 
pirical knowledge is probable only. (8) The mu- 
tual independence of the concept and the given, 
and the merely probable character of empirical 
truth, are entirely compatible with the validity of 
cognition. The problem of the "deduction of the 


categories" can be met without any metaphysical 
assumption of a preestablished amenability to 
categorial order in what is independent of the 
mind. (9) More explicitly, any conceivable ex- 
perience will be such that it can be subsumed un- 
der concepts, and also such that predictive judg- 
ments which are genuinely probable will hold of it. 
This chapter and the next are devoted to the 
distinction of the two elements in experience, and 
to the defense of this distinction from various 
common misinterpretations. 

There are, in our cognitive experience, two ele- 
ments ; the immediate data, such as those of sense, 
which are presented or given to the mind, and a 
form, construction, or interpretation, which rep- 
resents the activity of thought. Recognition of this 
fact is one of the oldest and most universal of 
philosophic insights. However, the manner in 
which these elements, and their relation to one an- 
other, are conceived, varies in the widest possible 
manner, and divergence on this point marks a 
principal distinction amongst theories of knowl- 
edge. As a result, even the most general attempt 
to designate these two elements as by the terms 
used above is likely to be objected to. Neverthe- 
less this distinction, in some terms or other, is ad- 
mitted to a place in almost every philosophy. To 
suppress it altogether, would be to betray obvious 
and fundamental characteristics of experience. If 


there be no datum given to the mind, then knowl- 
edge must be contentless and arbitrary; there 
would be nothing which it must be true to. And 
if there be no interpretation or construction which 
the mind itself imposes, then thought is rendered 
superfluous, the possibility of error becomes in- 
explicable, and the distinction of true and false is 
in danger of becoming meaningless. If the sig- 
nificance of knowledge should lie in the data of 
sense alone, without interpretation, then this sig- 
nificance would be assured by the mere presence 
of such data to the mind, and every cognitive ex- 
perience must be veracious. 

There are, to be sure, theories which emphasize 
one of these two elements almost to the exclusion 
of the other. Such theories are of both sorts 
those which emphasize what is given and those 
which emphasize the active mind. Immediacy is 
thus emphasized by the mystics generally, by 
Bergson, and by the American new-realists to 
mention only those examples which will come at 
once to the reader's mind. The idealists, on the 
other hand (empirical idealists, like Berkeley, ex- 
cepted), may seem to include the content as well 
as the form of knowledge in what the activity of 
thought creates. However, a closer examination of 
such theories, of both sorts, will usually reveal that 
the distinction is still recognized ; it is merely ob- 
scured by preoccupation with other issues. 

In theories of the first type, which identify 


knowledge with some state of pure Immediacy, the 
description or analysis of the cognitive experience 
is subordinated to the attempt to establish the 
superior value of some one type of experience as 
compared with others. The mystic, for example, 
values preeminently that experience which he in- 
terprets as being the immediate presence to, and 
coalescence with, his own mind of the transcendent 
object which he seeks. But he will readily grant 
the presence and determining character of con- 
ceptual interpretation in ordinary non-mystical 
experience. Only he condemns the object of such 
experience as illusion or mere appearance. The 
world of every-day is not, for him, ultimately real ; 
or at least its true nature is not revealed in ordi- 
nary experience. The moment of true insight is 
that in which the distinctions and relations which 
discursive thought creates are shorn away and 
reality stands forth, in luminous immediacy, as it 
truly is. Now all men restrict the word "knowl- 
edge" to the apprehension of the real. Hence the 
mystic's metaphysical conception, which leads him 
to use the word "real 55 differently than other men, 
likewise moves him to restrict the term "knowl- 
edge" to the peculiar experience in which this 
"reality" is apprehended. That in the ordinary 
experience which other men trust as truly cogni- 
tive, the element of interpretation is present, he 
fully recognizes and even insists upon. He recog- 
nizes also that this conceptual element represents 


something induced by the construction or attitude 
of the mind itself.* 

The reason why Bergson identifies the truest 
knowledge with "intuition" is similarly rooted in 
metaphysical theory and not in any divergent 
reading of our ordinary experience. For him, the 
ultimate reality is life, or the inwardly grasped 
"real duration.* 5 For each mind, this is something 
which is immediate, in his own case, and is to be 
apprehended in its other manifestations only by 
empathy or einfuhlung. The world of science and 
common sense Bergson recognizes to be construc- 
tion or interpretation which the mind imposes 
upon the data of immediacy. Also, he is explicit 
that this construction is dominated by interests of 
action and of social cooperation. But the space- 
world which results from such interpretation, he 
regards as not an ultimate reality ; hence the cog- 
nitive experience which includes this interpretive 
element is not a theoretically adequate knowledge. 
In short, with Bergson as with the mystics, identi- 
fication of knowledge with intuitive apprehension 

*Often the mystic, inheriting his terminology from Aristotle^ in- 
terprets the attitude of mind in every-day experience as passivity 
rather than activity, reserving the latter term for his own kind of 
absorbed concentration. For him, the interpretation which charac- 
terizes ordinary thought is at the behest of enslaving "passions.'* 
That such construction is significant of ordinary and mundane in- 
terests, he fully understands. But such interests are, for him, to be 
avoided and quelled. Here again, his use of terms, reversing the usual 
one, is governed by metaphysical and ethical preoccupation which 
is irrelevant to the just analysis of mundane experience. He reservee 
the laudatory term "active" for the ethical attitude which he seeks 
to inculcate. 

It is thus that in philosophy we give over the accurate report of 
fact to quarrel for exclusive possession of honorific terms. 


of the immediate reflects no basic difference in the 
analysis of ordinary experience but rather a dif- 
ference in the denotation given to the phrase "true 
knowledge" because of a metaphysical theory 
which denies ultimate reality to what is cognized 
by science and common sense. 

Of all the current theories in which knowledge 
is portrayed in terms of receptivity alone, the new 
realism would seem to be the only one in which 
this predilection for the given does not reflect a 
metaphysical preoccupation. Here the activity of 
thought (or attention) is represented as selective 
only; it may determine what is included in or ex- 
cluded from perception, but it does not supple- 
ment or modify the given data. Mind, so far as 
mind is just now a knowing of this object, and the 
object, so far as it is just now known by this 
mind, are represented as coinciding.* 

Any such theory must reveal its inadequacy by 
failure to account satisfactorily for the possibility 
of error. So far as knowledge is pure receptivity, 
that with which the mind coincides in cognition 
must in all cases have the same objectivity. Or at 
least, no ground is here provided for the distinc- 
tion between veridical and illusory apprehension. 
Thus we have the question whether mirror-images 
are truly located in the space to which they are 

*It would seem that most, or all, of those who cooperated in the 
volume, "The New Realism," have since abandoned or considerably 
modified the positions there taken, so that what is here said may be 
a discussion of nobody's present conception. But the theory is of in- 
terest on its own account. 


referred, the difficulty about the star seen now, 
though it may have ceased to send its light-rays 
from that point a thousand years ago, and a num- 
ber of like problems. The new realist may go the 
whole length, as Mr* Holt did, and hold that con- 
tradictories and incompatibles can be objectively 
real.* But in that case he ceases, for most of us to 
be plausible. Or he may, as Mr. Montague did, 
introduce some theory of a plurality of causes 
which can produce the same brain-state, and ex- 
plain error through the ambiguity thus intro- 
duced. But it would appear that, apart from any 
question about indentifying brain-states with 
perceptions, or any question about the propriety 
of the element of representationalisni thus intro- 
duced into what is otherwise a purely "presenta- 
tion" theory of knowledge, it will be impossible, 
on this account, to escape the admission of an ele- 
ment of interpretation in cognition. So long as 
the content of knowledge merely coincides with 
what is presented, knowledge must still be always 
veridical, because the brain-state (or perception) 
will contain only so much of the plurality of 
causes which may give rise to it as will in all cases 
coincide. The brain-state can be identical only 
with what is identical in the plurality of its causes 
unless we wish to abandon the principle that 
things identical with the same thing are identical 
with each other. If a single brain-state, or modi- 
*"The Concept of Consciousness/* 1914. 


fication of perceptive consciousness, is taken as 
meaning one thing when its veridical significance 
is of another, then some interpretation which goes 
beyond the content of this given state itself is the 
only conceivable basis of the error. 

Furthermore, it is impossible to escape the fact 
that knowledge has, in some fashion and to some 
degree, the significance of prediction. As Berkeley 
put it, one idea or presentation is sign of an- 
other which is to be expected. So far as this is 
true, the cognitive significance may attach to the 
data of sense but cannot simply coincide with such 
given data. To know is to find what is presented 
significant of what is not, just now, so presented. 
It is because Berkeley failed to follow out the im- 
plications of his own theory and to examine the 
validity of this relation, by thought, of what is 
given to what is not thus immediate, that the way 
lay open to Hume's skepticism. 

Failure to recognize and consider this element 
of construction or interpretation by the mind, will 
wreck any theory of knowledge. Failure to ac- 
knowledge its existence will make it impossible to 
account for error. And failure to find the ground 
of its validity will lead inevitably to skepticism; 
if not to skepticism ordinarily so-called, at least 
to that skepticism of every-day cognition which is 
involved in immediacy theories such as mysticism. 

With theories of the opposite sort, which em- 
phasize the constructive mind and seem to exclude 


any independent given such as sense-data, the 
explanation may similarly be found in a meta- 
physical preoccupation* This is obviously the case 
with Plato.* The data of sense are, for him, not 
relevant to true knowledge because only the tran- 
scendent ideas are fully real. In that mixed sort 
of apprehension by which the physical external 
world is grasped, the place of sense-data, on his 
account, is evident. 

Post-Kantian idealism also may seem to con- 
tend for the identification of knowledge with what 
the activity of thought alone produces. But ideal- 
ism can hardly mean to deny that the fact of my 
seeing at this moment a sheet of white paper in- 
stead of a green tree is a datum which it is beyond 
the power of my thought to alter. It can hardly 
mean to deny the given in every sense. As a fact, 
idealists of this school seldom speak directly to 
the question : "Does the activity of thinking create 
what would ordinarily be called the data of sense ?" 
This question may not seem to them important 
because their metaphysical thesis does not turn 
upon it, but rather upon two somewhat different 
issues : "Can there be any apprehension of a real 
object without the active construction of the 
mind?" and, "Is the existence of sense-data, as 
such, evidence of a reality which is independent of 

*It is not dear that Plato is activist, rather than intuitionist, in 
his conception of noesis, but at least sense-data have no part in this 
highest kind of knowledge. 

(See, for example, Green: ** Prolegomena to Ethics,** chap. I, sees. 
13 and 13. 


The first of these questions is answered to their 
satisfaction if it can be shown that the objectivity 
of the real requires always construction by the 
mind. This thesis does not imply any denial that 
the given is independent of the activity of thought 
in the sense explained above- It requires only the 
denial that the presentation of sense-data can by 
itself constitute valid knowledge- That I credit 
this presentation, or attribute to it objectivity, is 
a judgment, and as such an act of thought. (It 
would equally be an interpretation to discredit 
the presentation as merely subjective). This in- 
terpretive fiat is what Fichte stresses as the posit- 
ing of the "not-me." The data of sense, apart 
from such positing, are neither external reality 
nor explicit self. In immediacy, there is no separa- 
tion of subject and object. The givenness of im- 
mediate data is, thus, not the givenness of reality, 
and is not knowledge. Hence the idealist may in- 
sist that there is no (real) object without the 
creative activity of thought, without in the least 
meaning to deny that there is a datum prior to 
its being posited as real, a content judged which 
is given to the judgment. As a fact, however, he 
often slights this point in his anxiety to pass on 
and refute the implications, contrary to his meta- 
physical thesis, which are frequently drawn from 

Also it may seem to the idealist more important 
to point out that given data are already in mind 


than to inquire whether such data are created by 
thought. If both the data interpreted and the in- 
terpretation put upon them belong to the mind, 
then the real object, as cognized, may be repre- 
sented as in both its aspects mind-dependent ; and 
no argument to an independent reality can be 
drawn from the analysis of knowledge. Thus the 
idealist may fail to admit, or even to recognize 
explicitly, that there are given data of experience 
which, merely as such and not as objective reality 
or unreality, the activity of thought can neither 
create nor alter* 

It is also characteristic of idealism to point out 
that the moment of pure givenness is a fiction, and 
its data an "unreal abstraction. 55 There is no ap- 
prehension he will insist without construction; 
hence the distinction of subject and object, act 
and given, must be within thought, not between 
thought and an independent something thought 
about. This consideration is of more importance 
for us, and will be discussed in the next chapter, 
But it implies no denial of the givenness of sense- 
data. It contends only that the mental state which 
should be purely receptive and coincide with the 
given is a fiction an observation which is unac- 
ceptable only to the mystic and other protagonists 
of pure intuition. Whether there is the beginning 
of a fallacious train of reasoning in this stretch- 
ing of the term "thought 55 to cover the cognitive 
experience as a whole, we need not pause to in- 


quire. At least, the denial of the given is not ob- 
viously necessary to any of the characteristic 
theses of idealism. Indeed, an unqualified denial 
of this element in ordinary cognition is sufficient 
to put any theory beyond the pale of plausi- 

We may, then, fairly take it for granted, as 
something generally recognized, that there are in 
experience these two elements, something given 
and the interpretation or construction put upon 
it. But the very fact that the recognition of this 
is so general and of such long standing enforces 
the necessity of considering the distinction with 
care. Different significances have been assigned to 
it, both historically and in contemporary thought. 
Moreover, various metaphysical issues gather 
about it: What is the relation between the given 
and the real? How does mind construct or inter- 
pret? What is this mind which can interpret : does 
it transcend experience? If so, how can it be 
known? If not, how can it condition experience in 
general by its interpretation? Confronted with 
such a tangle of problems, we shall do best, I 
think, if first we can catch our facts. If we can 
identify the thing to be discussed, a certain degree 
of clarity will accrue simply by telling the truth 
about it, if we can. 

There is, in all experience, that element which 
we are aware that we do not create by thinking 
and cannot, in general, displace or alter* As a first 


approximation, we may designate it as "the sensu- 

At the moment, I have a fountain pen in my 
hand. When I so describe this item of my present 
experience, I make use of terms whose meaning 
I have learned. Correlatively I abstract this item 
from the total field of my present consciousness 
and relate it to what is not just now present in 
ways which I have learned and which reflect modes 
of action which I have acquired. It might happen 
that I remember my first experience of such a 
thing. If so, I should find that this sort of pres- 
entation did not then mean "fountain pen 35 to me* 
I bring to the present moment something which I 
did not then bring ; a relation of this to other ac- 
tual and possible experiences, and a classification 
of what is here presented with things which I did 
not then include in the same group. This present 
classification depends on that learned relation of 
this experience to other possible experience and 
to my action, which the shape, size, etc., of this 
object was not then a sign of. A savage in New 
Guinea, lacking certain interests and habits of 
action which are mine, would not so classify it. 
There is, to be sure, something in the character 
of this thing as a merely presented colligation of 
sense-qualities which is for me the clue to this 
classification or meaning; but that just this com- 
plex of qualities should be due to a **pen" char- 
acter of the object is something which has been 


acquired. Yet what I refer to as "the given" in 
this experience is, in broad terms, qualitatively no 
different than it would be if I were an infant or 
an ignorant savage. 

Again, suppose my present interest to be 
slightly altered. I might then describe this object 
which is in my hand as "a cylinder" or "hard 
rubber" or "a poor buy." In each case the thing 
is somewhat differently related in my mind, and 
the connoted modes of my possible behavior to- 
ward it, and my further experience of it, are dif- 
ferent. Something called "given" remains con- 
stant, but its character as sign, its classification, 
and its relation to other things and to action are 
differently taken. 

In whatever terms I describe this item of my 
experience, I shall not convey it merely as given, 
but shall supplement this by a meaning which has 
to do with relations, and particularly with rela- 
tion to other experiences which I regard as pos- 
sible but which are not just now actual. The man- 
ner of this supplementation reflects my habitual 
interests and modes of activity, the nature of my 
mind. The infant may see it much as I do, but 
still it will mean to him none of these things I 
have described it as being, but merely "play- 
thing" or "smooth biteable." But for any mind 
whatever, it will be more than what is merely 
given if it be noted at all. Some meaning of it also 
will be contained in the experience. All that comes 


under this broad term "meaning" (unless immedi- 
ate value or the specificity of sense-quality should 
be included) is brought to this experience by the 
mind, as is evidenced by the fact that in this re- 
spect the experience is alterable to my interest and 
my will. 

This meaning or interpretation or construction 
which is attached to the given is significant in two 
directions, connected but different. The one is the 
relation of this which is immediately presented to 
further actual and possible experience ; the other 
is its relation of my interest and action. The rela- 
tion to other experience, is something which is 
brought to the present by a selective memory. As 
applied to this present given, however, it is sig- 
nificant, not of the past, but of an actual or pos- 
sible future, continuous with this present moment. 
Thus this given is set in a relation with a to-be- 
given or could-be-given, and this setting is an in- 
terpretation of it which the temporal process of 
experience may verify or prove erroneous. The 
other relation of the given to present interest or 
attitude connotes an interplay between the tem- 
poral process of further possible experience and 
my own purposes and behavior. Since I not only 
think but physically act, I enter into the temporal 
process of the future as a factor which determines, 
in some part, what it shall present. Thus my in- 
terpretation is predictive of my own physical be- 
havior as forecast by my present interested atti- 


tude, and of further experience as affected by 
that behavior. In all those ways in which my in- 
terpretation could be phrased as predictive not 
of future actual but only of possible experience, 
it very likely has reference to ways of acting 
which I know I might adopt at will and the future 
experience which I should then expect. 

My designation of this thing as "pen" reflects 
my purpose to write; as "cylinder" my desire to 
explain a problem in geometry or mechanics; as 
"a poor buy 55 my resolution to be more careful 
hereafter in my expenditures. These divergent 
purposes are anticipatory of certain different fu- 
ture contingencies which are expected to accrue, 
in each case, partly as a result of my own action. 

The distinction between this element of inter- 
pretation and the given is emphasized by the fact 
that the latter is what remains unaltered, no mat- 
ter what our interests, no matter how we think or 
conceive. I can apprehend this thing as pen or 
rubber or cylinder, but I cannot, by taking 
thought, discover it as paper or soft or cubical. 

While we can thus isolate the element of the 
given by these criteria of its unalterability and 
its character as sensuous feel or quality, we cannot 
describe any particular given as such, because in 
describing it, in whatever fashion, we qualify it by 
bringing it under some category or other, select 
from it, emphasize aspects of it, and relate it in 
particular and avoidable ways. If there be states 


of pure esthesis, in violent emotion or in the pres- 
ence of great art, which are unqualified by 
thought, even these can be conveyed and perhaps 
even retained in memory only when they have 
been rendered articulate by thought. So that in 
a sense the given is ineffable, always. It is that 
which remains untouched and unaltered, however 
it is construed by thought. Yet no one but a phi- 
losopher could for a moment deny this immediate 
presence in consciousness of that which no activity 
of thought can create or alter. 

If now we have fastened upon the fact of ex- 
perience which we wish to discuss as the given 
element in it, it is time that we proceed to clarify 
this conception and guard against various pos- 
sible misinterpretations. 

An initial difficulty may arise from ambiguity 
of the word "given." This term has most fre- 
quently been used in philosophy in meanings at 
least close to the one here intended. But on occa- 
sion it has the widely different significance of de- 
noting those data which philosophy in general 
finds or takes for granted at the beginning of its 
study. And occasionally those who use the term in 
this second meaning make it carry something of 
methodological polemic against any notion that 
*'the immediate" or "sense-data" are allowable 
categories of explanation in epistemology. 

What I should have to say on this point is, in 
part at least, already clear from the preceding 


chapter. It is indeed the thick experience of the 
world of things, not the thin given of immediacy, 
which constitutes the datum for philosophic re- 
flection. We do not see patches of color, but trees 
and houses ; we hear, not indescribable sound, but 
voices and violins. What we most certainly know 
are objects and full-bodied facts about them 
which could be stated in propositions. Such initial 
data of object and fact set the problem in phi- 
losophy and are, in a measure, the criteria of its 
solution, since any philosophic theory will right- 
fully be rejected as inaccurate or inadequate if it 
does not measure up to, or account for, experience 
in this broad sense* 

But the acceptance of such preanalytic data* 
as an ultimate epistemological category would, if 
really adhered to, put an end to all worthwhile 
investigation of the nature of knowledge or to 
any other intellectual enterprise. What lies on the 
surface can be taken as ultimate only so long as 
there is no problem to be solved, or else no solu- 
tion to be hoped for. Without analysis, there can 
be no advance of understanding. 

The given, as here conceived, is certainly an ab- 
straction. Unless there be such a thing as pure 
esthesis (and I should join with the critic in 
doubting this), the given never exists in isolation 
in any experience or state of consciousness. Any 

*I borrow this useful pBrase from Professor Loewenberg; see has 
article, "Preanalytic and Postanalytic Data," The Journal of Philos- 
ophy, vol. 24 (1927), pp. 5 Jf. 


Kantian "manifold' 3 as a psychic datum or mo- 
ment of experience, is probably a fiction, and the 
assumption of it as such is a methodological error. 
The given is in, not before, experience. But the 
condemnation of abstractions is the condemnation 
of thought itself. Nothing that thought can ever 
comprise is other than some abstraction which 
cannot exist in isolation. Everything mentionable 
is an abstraction except the concrete universal; 
and the concrete universal is a myth. Thought 
can do just two things : it can separate, by analy- 
sis, entities which in their temporal or spatial ex- 
istence are not separated, and it can conjoin, by 
synthesis, entities which in their existence are dis- 
joined. Only the mystic or those who conceive that 
man would be better off without an upper-brain, 
have ground for objection to analysis and abstrac- 
tions. The only important question is whether this 
abstracted element, the "given," is genuinely to 
be discovered in experience. On this point I can, 
of course, only appeal to the reader. I shall hope 
that he has already identified provisionally what 
the word intends, and proceed upon that basis. 

Assuming, however, such provisional identifica- 
tion, there are still ambiguities of language to be 
avoided. I have so far spoken of "the given" and 
"data of sense" as roughly synonymous, but the 
latter phrase has connotations which are slightly 
inappropriate. In the first place, "sense-data," as 
a psychological category, may be distinguished 


from other mental content by their correlation 
with the processes in the afferent nerves. A dis- 
tinction made by the criterion of such correlation 
with nervous processes is open to two objections 
in epistemology. In the first place, the thing or 
mental state itself must first be accurately iden- 
tifiable before such correlation can be established. 
If it is thus identifiable, the correlation is not es- 
sential and is, in fact, superfluous in discussions 
of epistemology by the method of reflection and 
analysis. Second, there is the more general objec- 
tion that the theory of knowledge is a subject too 
fundamental to rest upon distinctions drawn from 
the particular sciences. Basic problems of cate- 
gory and of the general nature of knowledge are 
antecedent to the special sciences and cannot, 
therefore, legitimately depend upon their particu- 
lar findings. Especially is this important as re- 
gards psychology, a valid method for which is, at 
the present moment, a serious problem. The man- 
ner in which my own body is known to me, the 
subjectivity or objectivity of pleasantness or emo- 
tional tone, the validity of the correlation between 
mental states which I inspect directly only in my 
own case (if such is the fact) and nervous proc- 
esses which I can observe only in another organ- 
ism ; these are themselves problems which have at 
least an epistemological side. 

Also, the particular purposes which the psy- 
chologist has in mind in making his analysis of 


mental states may be out of place in epistemology. 
"Sense-datum" may connote relation to particu- 
lar sense-organs (as in the distinction between 
taste and odor), and hence mark a division where 
none can be drawn by direct inspection. Also 
other qualities than the strictly sensory may be 
as truly given ; the pleasantness or f earf ulness of 
a thing may be as un-get-overable as its bright- 
ness or loudness that question, at least, must not 
be prejudiced. Hence "sense-data," defined by 
correlation with nervous processes, should have 
no place in our program. It is the brute-fact ele- 
ment in perception, illusion and dream (without 
antecedent distinction) which is intended. 

However, if it be understood that the methodo- 
logical connotation of psychological categories is 
not here in point, it will cause no confusion if I 
continue to refer to the "sensuous" or the "feel- 
ing" character of the given: the element in ex- 
perience which is intended is difficult to designate 
in any terms which are not thus preempted to 
slightly different uses. It seems better to use lan- 
guage which is familiar, even at some risk of am- 
biguity, than to invent a technical jargon which 
would, after all, be no less ambiguous until its 
precise connotation could be discovered from its 

There is also another and different kind of am- 
biguity which must be avoided* Obviously, we must 
distinguish the given from the object which is 


given. The given is presentation of something 
real, in the normal case at least; what is given 
(given in part) is this real object. But the what- 
ness of this object involves its categorial inter- 
pretation; the real object, as known, is a con- 
struction put upon this experience of it, and 
includes much which is not, at the moment, given 
in the presentation. 

Still further comment is required in view of 
contemporary theories which deal with the con- 
tent of immediacy in a different fashion.* 

When we remember that even the delimitation 
of that in which we are interested, the singling 
out of the presentation of our object from other 
accompanying consciousness is, in some part at 
least, a work of excision or abstraction wrought 
by the mind, we may be led to remark that there 
is, in all strictness, only one given, the Bergson- 
ian real duration or the stream of consciousness. 
This, I take it, is at least approximately correct. 
The absolutely given is a specious present, fading 
into the past and growing into the future with no 
genuine boundaries. The breaking of this up into 
the presentation of things marks already the ac- 
tivity of an interested mind. On the other hand, 
we should beware of conceiving the given as a 
smooth undifferentiated flux; that would be wholly 

*Extended discussion of such theories cannot be included here; 
the object of the discussion is merely that of clarifying, by contrast, 
the terminology and procedure here adopted. Though criticisms will 
be ventured, it is recognized that the discussion is not sufficient fully 
to substantiate these. 


fictitious. Experience, when it comes, contains 
within it just those disjunctions which, when they 
are made explicit by our attention, mark the boun- 
daries of events, "experiences" and things* The 
manner in which a field of vision or a duration 
breaks into parts reflects our interested attitudes, 
but attention cannot mark disjunctions in an un- 
differentiated field* 

The interruptions and differences which form 
the boundaries of events and things are both given 
and constituted by interpretation. That the rug is 
on the floor or the thunder follows the flash, is as 
much given as the color of the rag or the loudness 
of the crash. But that I find this disjunction of 
rug and floor possessed of a meaning which the 
wrinkles in the rug do not have, reflects my past 
experience to taking up and putting down rugs. 
The cognitively significant on-the-floorness of the 
rug requires both the given break in the field of 
vision and the interpretation of it as the boundary 
between manipulable object and unyielding sup- 

Even in that sense in which the given is always 
one whole, it is not important f ot our purpose of 
analyzing knowledge that we should dwell upon 
this integrality of it. Our interest is, rather, in the 
element of givenness in what we may, for usual and 
commonplace reasons, mark off as "an experience" 
or "an object*' 5 This given element in a single ex- 
perience of an object is what will be meant by "a 


presentation/' Such a presentation is, obviously, 
an event and historically unique. But for most of 
the purposes of analyzing knowledge one presenta- 
tion of a half-dollar held at right angles to the line 
of vision, etc., will be as good as another. If, then, 
I speak of "the presentation" of this or that, it 
will be on the supposition that the reader can pro- 
vide his own illustration. No identification of the 
event itself with the repeatable content of it is in- 

In any presentation, this content is either a spe- 
cific quale (such as the immediacy of redness or 
loudness) or something analyzable into a complex 
of such. The presentation as an event is, of course, 
unique, but the qualia which make it up are not. 
They are recognizable from one to another expe- 
rience* Such specific qualia and repeatable com- 
plexes of them are nowadays sometimes designated 
as "essences/ 5 This term, with such a meaning, will 
here be avoided ; the liability to confuse such qualia 
with universal concepts makes this imperative. 

It is at once the plausibility and the fatal error 
of "critical realism" that it commits this confusion 
of the logical universal with given qualia of sense 
by denominating both of these "essences." As will 
be pointed out later, what any concept denotes 
or any adjective such as "red" or "round" is 
something more complex than an identifiable sense- 
quale. In particular, the object of the concept must 
always have a time-span which extends beyond the 


specious present ; this is essential to the cognitive 
significance of concepts. The qualia of sense as 
something given do not, in the nature of the case, 
have such temporal spread. Moreover, such qualia, 
though repeatable in experience and intrinsically 
recognizable, have no names. They are fundamen- 
tally different from the "universals" of logic and 
of traditional problems concerning these. Eluci- 
dation of this point must wait upon the sequel. 

The somewhat similar use of the terms "sensa" 
and "sense-data" is also likely to prove preju- 
dicial. Mr. Broad in particular has used these 
terms in a fashion which gives what is denoted by 
them a dubious metaphysical status. He says, for 
example:* "We agreed that, if they (sensa) are 
states of mind at all, they must be presentations. 
But we find no reason for thinking that they are 
states of mind, and much the same reasons against 
that view as led us to hold that sensations are 
analyzable into act and sensum. * . . We saw no 
intrinsic reason why coloured patches or noises 
should not be capable of existing unsensed." And 
elsewhere :f "A sense-datum with which I am ac- 
quainted may perfectly well have parts with which 
I am not acquainted. If therefore I say that a 
given sense-datum has no parts except those which 
I have noticed and mentioned I may quite well be 
wrong. Similarly there may well be differences of 

*" Scientific Thought," p. 265. 

t" Proceedings of the Aristotelian Society," supp. vol. IE, p. 18. 


quality which. I cannot detect. If I say : This sense- 
datum with which I am acquainted is coloured all 
over with a uniform shade of red, this statement 
may be false/ 5 

Now it is indeed obvious that I may make 
erroneous report of the given, because I can make 
no report at all except by the use of language, 
which imports concepts which are not given. The 
sensum-theory, like the essence-theory, fails to go 
deep enough and to distinguish what is really 
given from what is imported by interpretation. 
There is interpretation involved in calling the 
sensum "elliptical" as much as in calling the 
penny "round" It means, for example, to assert 
something about the motion one must make with 
one's finger in order to hide successively the dif- 
ferent portions of the periphery. Also it is true, 
of course, that if I report the given as "red," I 
may convey an erroneous impression because I am 
heedless of color-meanings, and another observer 
who should have an experience qualitatively the 
same might report "orange" or "violet." Similarly 
I may report "round" when an artist would re- 
port "elliptical," because I am not used to pro- 
jecting things on a plane. All those difficulties 
which the psychologist encounters in dealing with 
reports of introspection may be sources of error 
in any report of the given. It may require care- 
ful self -questioning, or questioning by another, to 
elicit the full and correct account of a given ex- 


perience. But Mr. Broad seems here to assert an 
entirely different ground of possible mistake. He 
seems to mean that with the same sensum before 
me I may at one moment see it red and at a later 
moment somewhat mottled or more deeply shaded 
at the center, and so forth. 

Now if I look fixedly at a card and see it first 
uniform and then mottled, I shall very likely and 
quite properly report that the color of the sur- 
face has a quality which I did not at first see. But 
the subject of this statement is the real color of 
the real card, and the statement itself is not a re- 
port of the content of sense but an interpretation 
put upon my succession of sensory experience. It 
imports a distinction between the subjective and 
the objective which is irrelevant to givenness as 
such. There certainly is such a thing as the shape 
of a penny or the color of a card which can exist 
unnoticed while I am looking at it or when I am 
not. This is because the shape of the penny has 
the same kind of enduring reality as the penny, 
and a quite different kind of reality from the in- 
termittent presentation of the penny in my con- 
sciousness. But I thought it was the initial point 
of the sensum-theory to provide a name if not a 
local habitation for what I see as opposed to 
what I see, the elliplicity of the appearance as 
against the real roundness of the penny. As such, 
it should be of the essence of the sensum to be 
sensed ; the sensum which is neither the real shape 


of the real penny nor the appearance of it in a 
mind is neither fish, flesh nor good red herring. A 
sensum which is not sensed, or a sense-datum 
which continues unaltered while consciousness of 
it changes, is merely a new kind of ding an $ic~h 9 
which is none the better for being inappropriately 
named so as to suggest its phenomenological char- 

What is given may exist outside the mind 
that question should not be prejudiced. But in or- 
der that we may meaningfully assert such exist- 
ence, it is essential that there be an answer to the 
questions: What would it mean if that which is 
given has such independent existence? In what re- 
spect would experience in general be different if it 
had no such independent reality? For a sensum 
which is not sensed, it is difficult to see what an- 
swer there can be to such questions.* The main 
objection to the sensum-theory is that it leaves at 

*I am never quite sure that I may not be misunderstanding Mr. 
Broad on this point. (I might say the same thing about Mr. Russell's 
"perspectives.") It may be that he means only that when no eye 
is situated at a certain angle from the penny, it is still true that if 
there were a normal observer there, he would observe this elliptical 
appearance* and that the appearance is there whether the observer 
is or not. To this I can give real meaning: the last part of the state- 
ment mptang the first part that the appearance is there when not 
observed, means that if any observer were there he would observe it. 
Or it means that other effects of the penny are there, such as an 
image registered on a camera-plate, and that these may be verified 
by the same methods as the existence of physical objects when not 
observed. But if what Mr. Broad means to call to our attention is 
the fact that the truth of the statement, "If an observer were there, 
he would observe so and so," can not be tested when no observer is 
there; and if its being thus true when not verified is what is meant 
by the existence of unsensed sensa; then I am compelled to say that 
the hypothesis is merely verbal nonsense. A hypothesis which in the 
nature of the case is incapable of any conceivable test is the hypothe- 


once the ground of the analysis of experience and 
plunges into metaphysics* It would explain the 
immediate and indubitable by something intrin- 
sically unverifiable and highly dubious. 

It is of the essence of what will here be meant 
by "the given 5 ' that it should be given. We need 
not say that what is given is a "mental state" or 
even "in the mind" in any more explicit sense 
than is itself implied in such givenness. Nor should 
it be presumed that what is thus in mind is e&- 
clusvvely mental. The nature of that interpreta- 
tion or construction by which we come to know ob- 
jects suggests that the given must be, in some 
sense or other, a constituent of objective reality 
as well. All such questions are simply later ques- 
tions. If there should be metaphysical problems 
concerning the kind of reality which what is "in 
mind" may have, it is not necessary to anticipate 
the solution of these beyond what may be verified 
in the discovery that certain items or aspects of 

sis of nothing. What is normally meant by saying, "If an observer 
were there he would observe so and so," is verifiable by the fact that, 
other conditions being altered at will, whenever an observer is there 
he does see this. As will be pointed out later, the attribution of prop- 
erties to objects and of existence to objects, consists, from the point 
of view of cognition, precisely in the truth of such hypothetical propo- 
sitions. And these are held to be true when the hypothesis is contrary 
to fact. But what we mean by the truth of such propositions is pre- 
cisely the sort of thing set forth above: the hjnpothetical, "If X were, 
then Y would be," means, "However other conditions be varied, 
and condition X being similarly supplied at will, whenever X is, Y 
is." If the existence of an appearance or sensum when it is not sensed 
means its observability at will, then it means its existence whenever 
it i$ sensed. So far as I can see, there is nothing else which such ex- 
istence reasonably could mean. In this, of course, I am merely ar- 
guing for the indispensability to meaning of the "pragmatic test" 
of Peirce and James. 


the content of experience satisfy the criteria of 
givenness. These are, first, its specific sensuous or 
feeling-character, and second, that the mode of 
thought can neither create nor alter it that it 
remains unaffected by any change of mental atti- 
tude or interest. It is the second of these criteria 
which is definitive ; the first alone is not sufficient, 
for reasons which will appear. 

This given element is never, presumably, to be 
discovered in isolation. If the content of percep- 
tion is first given and then, in a later moment, in- 
terpreted, we have no consciousness of such a first 
state of intuition unqualified by thought, though 
we do observe alteration and extension of inter- 
pretation of given content as a psychological tem- 
poral process. A state of intuition utterly unquali- 
fied by thought is a figment of the metaphysical 
imagination, satisfactory only to those who are 
willing to substitute a dubious hypothesis for the 
analysis of knowledge as we find it. The given is 
admittedly an excised element or abstraction; all 
that is here claimed is that it is not an "unreal" 
abstraction, but an identifiable constituent in ex- 


We have so far been concerned mainly with the 
distinction between the two elements in knowl- 
edge, the given and the construction or interpreta- 
tion put upon it, and particularly with the criteria 
of the given. We turn now to consideration of the 
conceptual or interpretational element. 

The word "concept" is used, in philosophic dis- 
cussion, in many different senses, three of which 
it is particularly important to distinguish. It may 
signify (1) the psychological state of mind when 
one uses a word or phrase to designate some indi- 
vidual thing or some class of objects. Or (2) it 
may refer to the meaning of a word or phrase 
throughout some period of the development of the 
individual's thought, or some period of the devel- 
opment of a science, of a given culture, or even 
of humanity altogether. Or, (3) it may signify 
the logical intension or connotation of a term. 
This third meaning is exemplified by dictionary 
definitions where these are satisfactory and is the 
usual signification of "concept** in the study of 

The use of any substantive phrase or term or- 
dinarily undergoes a process of development, both 
in the history of society and in the history of any 



individual who uses it. Usually, though not al- 
ways, the denotation of the term remains un- 
changed throughout this process; we apply it to 
the same class of objects, but our realization of 
what is essential to these things reflects a process 
of learning. Such learning may consist in an en- 
largement of our experience of the class of things 
in question, or it may occasionally represent sim- 
ply our more accurate apprehension of what are 
the universal properties and relations of the fa- 
miliar objects thus classified. But if the meaning 
of a word or phrase undergo evolution, then, how- 
ever normal or inevitable or commendable this 
process may be, we must, for the sake of clarity, 
recognize that this meaning is one unitary entity 
only in some generic and genetic sense, and that 
logically what we have is a succession of different 
meanings, related in ways which may be impor- 
tant. The recognition of their historical conti- 
nuity must not obscure the fact of their logical 

The problem of the developing adequacy of 
thought is an interesting and important one; it 
requires as one of its fundamental categories the 
notion of just such historical and psychological 
continuities; and the selection of the word "con- 
cept" to designate this category is natural. No 
criticism of such genetic study or of this use of 
the term is intended. But this psychological and 
educational category must not be confused with 


"meaning 55 in the sense in which logic, for in- 
stance, requires that term. Here a meaning must 
be precise and clear, or be capable of being made 
so, and must remain unaltered throughout any 
discussion in which it occurs. No psychological or 
historical process is legislated out of existence by 
this restriction in the use of the word, but if there 
should be development or learning which affects 
the connotation of a term, then, from this point of 
view, we have another meaning; that is all. 

Again, the psychological state is not the object 
in which we are here primarily interested. If a 
psychologist, thinking in terms of a context-the- 
ory of meaning, says, "Infinity means to me the 
image of the blue-black, dense, arched sky,' 5 * then 
we must observe that such a psychologist blurs 
over the distinction between what is essential and 
what is non-essential in meaning. He is in no dan- 
ger of misunderstanding one who talks about 
what the symbol CO denotes to be referring to the 
heavens, nor does he, even in his own thinking, 
suppose that infinity is blue-black. To use "con- 
cept 55 to designate such a psychological state or 
association-complex, is to fail to mark the dis- 
tinction between what is objective in meaning and 
what is adventitious or purely personal. Indeed, 
the question how meaning can be objective and 
shared, when the psychological states which are 
bearers of this meaning are separate existences 

*The reference is, of course, to Titchener. 


and not even identical in their qualitative content, 
is one of the important problems of meaning. 

Because it is our main interest here to isolate 
that element in knowledge which we can with cer- 
tainty maintain to be objective and impersonal, 
we shall define the pure concept as "that meaning 
which must be common to two minds when they 
understand each other by the use of a substantive 
or its equivalent/ 5 (For brevity, the qualification 
"pure" will be omitted throughout the remainder 
of this chapter.) However, this designation of 
community of meaning as the distinguishing mark 
of the concept is, in part, merely an expository 
device for singling out that element in knowledge 
which, for reasons which will appear, I wish here 
to discuss. 

That meanings may have this sort of objectiv- 
ity, is a fundamental assumption of science or of 
any other intellectual enterprise. If there is noth- 
ing objective about propositions and concepts, 
then there is no such thing as truth and there can 
be no serious purpose in reflection or discussion. 
There must be meanings which are common to 
minds when they cooperate in scientific or even in 
merely practical endeavors. Otherwise the cooper- 
ation is illusory; and one cannot escape the ques- 
tion how such common meaning stands related to 
different minds or psychological states which 
mean. One may follow Plato and cut the Gordian 
knot by removing these precise and logically mean- 


ings beyond our earthly sphere and establishing 
them as transcendent ideas or eternal objects. 
This reflects a judgment of their value but leaves 
our commerce with them a miracle ; it substitutes 
adoration of a mystery for explanation of a fact. 
A similar remark would apply to any doctrine like 
that of the new realism, so far as this doctrine 
hypostatizes conceptual realities, such as those of 
mathematics, setting them up as objective reali- 
ties without further ado and then explaining our 
apprehension of them as coincidence of mind and 
object. One does not answer the numerous objec- 
tions which the nominalists and conceptualists in 
logic have urged and very plausibly urged by 
first setting up what they claim can exist only in 
a mind as something outside it and then offering 
coincidence of mind and object as explanation of 
the fact that these conceptual objects are also m 
minds. The new realist here follows the obvious 
analogy of the common-sense view concerning 
knowledge of the physical. There is the brick out 
there; we both see the brick, hence we have an 
idea in common. So the new realist seems to say 
there is the mathematical entity out there; we 
both apprehend this mathematical entity, hence 
we have an abstract mathematical concept in com- 
mon. Even in the case of the physical object there 
are all sorts of difficulties to be met before com- 
munity of knowledge can be understood. And in 
the case of the purely abstract or conceptual (if 


there be any purely conceptual entities) we have 
the added difficulty that such an object cannot 
arouse sensation. Undoubtedly the conceptual has 
its own appropriate kind of reality ; but what that 
kind of reality is, is precisely the problem. It is 
not to be resolved by a phrase such as "neutral 
entity/ 5 or "eternal object," or "essence/' We 
must assume the objectivity of conceptual mean- 
ing. But if in order to philosophize sensibly we 
must assume something to be true when we do not 
understand how it can be true, then our philoso- 
phy is, so far, a failure. 

On the other hand, I see no necessity for resign- 
ing the problem of common meanings to the psy- 
chologist as his exclusive affair, especially since he, 
like the rest of us, must begin by assuming their 
existence* The meaning must be somehow identi- 
fied before it can be correlated with behavior or 
motor-set or context or anything else. And it 
must be identified as somehow common to two 
minds before individual psychological differences 
will be pertinent to it ; if what they are pertinent 
to is not somehow identical, but A*s state is per- 
tinent of sc and J5's to #, then there is no basis 
for comparison by which individual differences 
could be discovered as such. 

Psychological differences of individuals are in- 
deed impressive. Long before scientific psychology 
was thought of, the skeptic appealed to them to 
prove the impossibility of knowledge or the com- 


munication of ideas. For imagery and feeling, and 
even to some extent for sensation, idiosyncrasy is 
the rule. Furthermore, as the ancient skeptic was 
fond of pointing out, there can be no final veri- 
fication of any community in these respects. The 
sense-quality of green cannot be conveyed to the 
congenitally blind; and if I suppose some idio- 
syncrasy of sense which makes my perception of 
green unique, I shall never discover that peculiar- 
ity provided it does not impair my powers to dis- 
criminate and relate as others do. In brief, there 
can be no verification of community between minds 
so far as it is a question of the feeling side of ex- 
perience, though the assumption that there is no 
community here seems fantastic. 

However, it is obvious that common meanings 
do transcend such individual differences of per- 
ception and imagery. We use language to convey 
thought. If language really conveys anything, 
then there must be something which is identical in 
your mind and in mine when we understand each 
other* And if our thought is objective and not 
merely a report of introspection, then what is 
identical in our two minds must also be somehow 
germane to that objective reality as we know it* 

Suppose we talk of physical things in physical 
terms, and our discussion involves physical mea- 
surement. Presumably we have the same ideas of 
feet and pounds and seconds* If not, the thing is 
hopeless. But in psychological terms, my notion of 


a foot goes back to some immediate image of 
visual so-long-ness, or the movements which I 
make when I put my hands so far apart, or a rela- 
tion between these two. Distances in general mean 
quite complex relationships between such visual 
images, muscle and contact-sensations, the feeling 
of fatigue, and so on. Weight goes back to muscle- 
sensations, the "heft" of the thing. And our direct 
apprehension of time is that feeling of duration 
which is so familiar but so difficult to describe. 

Now in such terms, will your sensory image of 
a foot or a pound coincide with mine? I am near- 
sighted; your eyes are good. My arms are long; 
yours are short. If we lift a weight, there is the 
difference in strength between us to take into ac- 
count* So it is with everything. In acuity of per- 
ception and power to discriminate, there is almost 
always some small difference between the senses 
of two individuals, and frequently these discrep- 
ancies are marked. It is only in rough and ready 
terms that we can reasonably suppose that our di- 
rect intuitions and images are alike. That so often 
theories of knowledge have ignored such differ- 
ences, which are the rule and not the exception, or 
have proceeded as if our common and supposedly 
veridical knowledge depended on coincidence of 
such sensory content, is really a frightful scandal. 

Even for the large and crude distinctions, what 
assurance is there that our impressions coincide? 
Suppose it should be a fact that I get the sensa- 


tion you signalize by saying "red" whenever I 
look at what you call "green" and vice versa. 
Suppose that in the matter of immediate sense- 
qualities my whole spectrum should be exactly the 
reverse of yours. Suppose even that what are for 
you sensations of pitch, mediated by the ear, were 
identical with my feelings of color-quality, medi- 
ated by the eye. Since no one can look directly 
into another's mind, and the immediate feeling of 
red or of the middle C can never be conveyed, 
how should we find it out if such personal pecu- 
liarities should exist? We could never discover 
them so long as they did not impair the power to 
discriminate and relate as others do. 

Furthermore, what difference to our common 
knowledge would it make? That is precisely the 
first point which needs to be emphasized: idio- 
syncrasy of intuition need not make any differ- 
ence, except in the esthetic quality of the experi- 
ence of one as compared with that of another. Let 
us take it for granted (it seems fairly sensible) 
that the sense-data of one are seldom precisely 
those of the other when we address ourselves to 
the same object. That, by itself, will in no way 
impede our common knowledge or the conveying 
of ideas. Why? Because we shall still agree that 
there are three feet to the yard; that yellow is 
lighter than blue; and that middle C means a 
vibration of 256 per second. In other words, if we 
define any one of the unit-ideas or concepts which 


enter into the expression of our thought, we shall 
define it in the same or equivalent fashion, and we 
shall apply the same substantives and adjectives 
to the same objects. So far as we fail in these two 
things, our knowledge will be really different and 
the attempt to convey our thought will break 
down. These, then, are the only practical and ap- 
plicable criteria of common knowledge; that we 
should share common definitions of the terms we 
use, and that we should apply these terms iden- 
tically to what is presented. 

I am not, of course, trying to argue that indi- 
vidual feelings are thus unique. Some differences 
of subjective experience are attested by the inabil- 
ity of one person to discriminate where another 
can. For the rest, the question of such identity is, 
in the end, mere idle speculation because we have 
no possible means of investigating it. What I 
would point out is, rather, that in the determina- 
tion of common concepts, the conveying of ideas, 
such possible idiosyncrasy in the correlated sense- 
feelings is entirely negligible. You and I mean 
the same by "red" if we both define it as the first 
band in the sun's spectrum, and if we both pro- 
nounce the same presented objects to be red. It 
does not matter if neither the red rug nor the 
first band of the spectrum give to the two of us 
identical sensations so long as we individually dis- 
cover that same sense-quality in each thing which 
we agree in describing as "red." 


Moreover, it is obvious that unless one have 
some peculiarity which both he and others will 
learn to recognize as a defect of his sense-percep- 
tions, the very manner in which we learn the 
names of things will secure such unanimity in the 
ascription of terms, regardless of any idiosyncrasy 
of purely inner sense feelings. 

Even those individual peculiarities which be- 
come recognized as inability to discriminate, limi- 
tation of the range of sensation, and so on, do not 
prevent us from sharing and conveying ideas, 
though they may impede the process of learning. 
We talk together and cooperate successfully about 
the vibration of 19,000 per second in the vacuum- 
tube, though for one of us this vibration is evi- 
denced by a note and for the other it never can 
be. We both have a perfectly good concept of 
ultra-violet, though neither of us will ever see it, 
just as we know well enough what we mean by the 
other side of the moon. To be sure, no such con- 
cept would have a meaning if we could not, 
through the terms in which that meaning is ex- 
plicated, get back eventually to concepts which 
are correlated for us with specific and identifiable 
qualities of sense. It is thus that we surmount our 
individual limitations. The pitch which is beyond 
my auditory range, I understand through the no- 
tion "vibration of 19,000," which is definitive both 
for me and for those who hear it as a note. This 
process of leading back, by which we understand 


what we can not directly perceive, may be quite 
complex and prolonged without defeating the pur- 
pose of sharing ideas and conveying thought* It 
is the same sort of process by which we must all of 
us understand what we mean by "ultr a- violet" or 
"electron," and its objectivity is not affected by 
the fact that such indirection of understanding is, 
in this case of limitations of perception, necessary 
for some of us only. 

The methods of verifying community of mean- 
ing are principally two, neither of which depends 
on any supposed community of feeling or imagery. 
Either we define our terms or, by our behavior,, 
we exhibit their denotation. The second proce- 
dure is less conclusive for reasons which are fairly 
obvious. No collection of cases, or examples 
pointed to, is ever sufficient to determine uniquely 
the denotation of a term to determine what other 
cases, not so far examined, would be included and 
what excluded. The meaning of the term will be 
what is common to the various examples pointed 
out as meant by it. In general, the larger the num- 
ber of things so indicated, the smaller the chance 
that these will have in common other properties 
than those which are essential or comprehended 
within the conceptual meaning of the term. But 
that possibility can never be ruled out. Moreover, 
the exhibition of meaning in this way depends for 
success upon the ability of the person to whom the 
meaning is conveyed to make the analysis which 


will isolate correctly just tliat totality of proper- 
ties which is common to the things indicated, not 
omitting to remark any which are essential. This 
is an important consideration, because concepts 
which stand in any need of being learned will rep- 
resent analyses which are matters of some diffi- 
culty. On account of these shortcomings of it, the 
actual use of this method of indicating a mean- 
ing by exemplifying its denotation is confined 
almost exclusively to conveying the meaning of a 
word where the concept itself is already something 
shared ; as, for instance, where teacher and taught 
have no language in common* 

The method of definition specifies a meaning 
directly* In defining, we refer one concept to oth- 
ers; the typical definition equates the term de- 
fined with a complex of other terms in definite 
relations. To be sure, it may not be sufficient that 
you and I both define A in terms of B and C, since 
B and C may have for us different significations. 
But if B also is defined by both in the same or 
equivalent fashion, and C 5 and so on, that is suffi- 
cient to assure common meaning, regardless of all 
differences of imagery or any idiosyncrasy of 
sense. Such verification of community of meaning 
by comparison of definitions is obviously a proc- 
ess which must be incomplete^ but it makes clear 
precisely what is essential to a genuine identity 
of meaning in two minds. 

Speaking in terms of logic, these facts may be 


expressed by saying that sensation and imagery 
are essentially individual, and do not possess 
meaning in the sense in which meanings are com- 
mon and shareable and expressible. The concept, 
which is the common, shareable and expressible 
meaning, must be distinguished from such feel- 
ings ; it is constituted by that pattern which is set 
up by the expression of one concept in terms of 
others. These patterns must be identical and 
shared by all who really possess ideas in common 
and are capable of conveying them to one another. 
Psychologically, this conceptual pattern of re- 
lations is, of course, an abstraction; no such con- 
cept ever existed, apart from imagery and sensory 
material, in any human mind. For each individual 
there must be a correlation of concept with spe- 
cific sense-quality. But this correlation of con- 
cept and sense is intrinsically individual; if it, 
too, should be shared, we could not verify that 
fact, and it is not in the least essential to com 1 
mon understanding that it should be. The con- 
cept, so defined, is precisely that abstraction which 
it is necessary to make if we are to discover the 
basis of our common understanding of that reality 
which we all know. On a day which is terribly long 
to me and abominably short to you, we meet, by 
agreement, at three o'clock, and thus demonstrate 
that we have a world in common. An "hour" is not 
a feeling of tedium or vivacity, but sixty minutes, 
one round of the clock, a pattern of relations 


which we have established between chronometers 
and distances and rates of movement, and so on. 

Defining, like logical analysis in general, sets 
up a pattern of relationships. We are all of us 
fond of what Bosanquet called the "linear" mode 
of thinking in such matters, and we might easily 
suppose that definition chases a conceptual mean- 
ing back into other such concepts, and these into 
still others, until finally it is brought to bay in 
some first (or last) identity of meaning which is 
identity of sensation or imagery. So far as mean- 
ing within the individual mind is concerned, I 
should suppose this is precisely what takes place ; 
we analyze the meaning back until we come to rest 
in familiar imagery. But the end-terms, which for 
us are thus understood directly by reference to 
sense and feeling, have still a conceptual meaning; 
they are not indefinable. This conceptual mean- 
ing is shareable; our imagery essentially not* 
Thus the end-terms of such analysis are no dif- 
ferent than the beginning terms ; they have mean- 
ing in two senses the logical, shareable meaning 
of further conceptual relations, and the direct, 
non-shareable meaning of reference to some com- 
plex of sense-qualities. 

The notion that the analysis of meaning must, 
in linear fashion, go back eventually to ultimate 
constituents whose meaning cannot in turn be 
thus relational, is a prejudice which is very largely 
due to a false metaphor. Logical analysis is con- 


ceived after the fashion of the physical dissection 
of a whole into parts, or the chemical analysis of 
a compound into elements* But it will not escape 
the thoughtful reader that all definition is eventu- 
ally circular. It is often the case that A can be 
defined in terms of B and C 9 B in terms of A and 
C, or C in terms of A and B. Where the circle is 
so small, and the defined meaning so promptly re- 
turns upon itself, the analysis is likely to be in- 
adequate. But this circularity would never be pos- 
sible at all, if the relation of defining to defined 
were that of part to whole* Moreover, the differ- 
ence between a good and a bad definition, on this 
point, is only, so to speak, in the diameter of the 
circle. All the terms in the dictionary, however 
ideal its definitions, will be themselves defined. 

Logical analysis is not dissection but relation; 
the analysis of A into B and C does not divide A 
into constituents B and C but merely traces a pat- 
tern of relations connecting A with B and C. As 
regards their conceptual meaning, terms are very 
closely analogous to points in space. A point is 
nothing whatever apart from its relation to other 
points ; its very essence is relational. Likewise the 
conceptual meaning of a term is nothing whatever 
apart from other such meanings. Also it is true 
that if point A is located by reference to B and 
C, B and C in turn, and the other points in any 
spatial array, have their position eventually, in 
circularwise, in their relation to A and to one 


another. The positional relationships of any point 
are internal to its nature and constitute that na- 
ture. Likewise, the definitive relations of a term, 
signifying a concept, are internal to the meaning 
of that term and constitute it. The nature of a 
concept as such is its internal (essential or defini- 
tive) relationships with other concepts. All points 
have their positions eventually in terms of the 
array of all space: no point or set of points has 
any primal position in any other fashion; we 
merely choose as an arbitary basis of reference 
some set which is convenient or marks the place 
where we happen to be. All terms or concepts 
similarly have their meaning eventually in the 
array of all meanings, and no member of this 
array is intrinsically primal or privileged* 

Concerning this interpretation of the concept 
as consisting in relational structures of meaning, 
there can be two doubts. We seldom "have in 
mind" any such conceptual pattern of definition. 
When we reflect upon the manner in which coin- 
cidence in the meaning of one term involves coin- 
cidence in the meaning of others, we see that such 
an ideal pattern of meaning goes far beyond what 
anyone could consciously have in mind at any one 
time. Again, we often coincide in our use of terms, 
and thus seem to possess meanings in common, 
when the definition of our terms would be a mat- 
ter of some doubt and one holding possibility of 


Three points are here pertinent: First, that 
over and above the ambiguities of language com- 
monly recognized, the same word may convey dif- 
ferent concepts on different occasions ; in particu- 
lar, that it may convey a meaning which is more 
or less restricted. Second, when the denotation of 
a word rules, there are degrees of clearness about 
its meaning. Third, identity of meaning consists 
practically in implicit modes of behavior, and what 
is involved in these always runs beyond what can 
be explicit in consciousness at any one time. 

If I talk with a chemist about helium or with a 
biologist about cells, we may understand each 
other perfectly. But without recourse to some ref- 
erence book, I could not define "helium" or "cell" 
in a fashion which the specialist would accept as 
adequate. To me "helium" means "a non-inflam- 
mable gas a little lighter than hydrogen (or a lit- 
tle heavier I forget which) , produced in the dis- 
integration of alpha-particles and found in the 
sun." I could not specify either atomic-weight or 
spectrum characteristics, one or other of which 
the chemist will regard as essential to a sufficiently 
guarded definition. But as long as we converse to- 
gether without misunderstanding, the common 
meaning of "helium" is just what is set down 
above. This is a less specific meaning than the 
chemist's, but included in it, and sufficient for our 
present purposes. If our discussion should touch 
upon more recondite matters, he might need to in- 


struct me about helium, and thus establish a more 
specific common concept, before we could go on* I 
recognize his authority to do so, and should accept 
his definition (which I cannot now give) as the 
"true" meaning of the word "helium." But this 
does not alter the fact that, for the time being, the 
common concept which serves our purposes is my 
looser understanding of the term. Such is quite 
commonly the case. Our actual meanings, the con- 
cepts we are concerned to convey, are more re- 
stricted than the true or full or dictionary mean- 
ings of the terms we use. Most words may convey 
any one of a whole range of more or less full and 
accurate meanings. It Is, thus 5 quite possible that 
we may understand each other perfectly even when 
we should disagree about the definition of our 
terms, because only some restricted meaning, cov- 
ered in both our definitions, is required by our 

Second, it is obvious that in some sense or other 
we may have a meaning in mind when we could 
not state it without further thought. Any true ac- 
count of thought and speech must recognize this. 
The ruling interest in knowledge is the practical 
interest of action. A meaning may be implicitly 
present in the consistency of behavior when con- 
fronted with experience of a certain type without 
the explicit recognition of what this behavior im- 
plies having come Into consciousness or even been 
brought in question. Such we must suppose to be 


the child's early use of language. And in this sense* 
we may perhaps say that meanings must be im- 
plicit in the mind before they can become con- 
scious. In fact, we may doubt whether any mean- 
ing would ever become conscious if it were not for 
the practical difficulties which arise when meanings 
are not thus explicit the difficulties of hesitant 
or inconsistent behavior in border-line cases, and 
the social difficulty of misunderstanding, that is, 
of incongruous behavior when the same term has 
been used with apparently the same meaning. 

Josiah Royce used to speak to his classes of the 
three grades of clearness about the meanings of 
terms.* We have the first grade of clearness when 
we are able appropriately to accept or reject any 
object of our acquaintance as belonging or not 
belonging to the class in question. The second 
grade of clearness involves, further, the prepared- 
ness to classify correctly objects not precisely like 
those with which we have previously been ac- 
quainted; that is, to make the dichotomy, X or 
not--3T, not only for familiar but also for unfa- 
miliar things, not only for all actual but also for 
all conceivable objects. The third grade of clear- 
ness consists in the ability to specify the criteria 
by which such classification is determined. This 
last, of course, is equivalent to definition, the ex- 
plicit possession of the concept. That the mind 

m *He used to attribute this to Charles Peirce, but Peirce's discus- 
sion m "How to Make Our Ideas Clear/* does not so precisely cover 
the point. 


may have the first or second grade of clearness 
without the third, is obvious. It is also evident 
without discussion, that even when we have, in the 
ordinary sense, this highest grade of clearness, we 
do not have this definition explicitly in mind when- 
ever we use a term with understanding. 

Any controversy as to whether a mind possesses 
a meaning whenever a term is used intelligently, 
would be useless because it would be verbal. The 
pertinent facts are sufficiently clear; that it may 
possess meaning in the sense of determining a con- 
sistent mode of behavior (such as the consistent 
use of a term) without our being able out of hand 
to specify the ground of our own discrimination, 
we can all of us testify* The psychology of this is 
doubtless a difficult and important topic ; but with 
that we are not concerned. It would be an anoma- 
lous use of language to deny meaning to terms 
which are used without this explicit consciousness 
of what is essential, especially since the use of 
terms, like other modes of deliberate behavior, is 
most frequently a matter of habit, reflecting pre- 
vious experience in which the mode of action was 
determined by clearer consciousness. It would also 
be anomalous to deny meaning where there is con- 
sistency of behavior or of consciously determined 
attitude which does not directly concern the use of 
language. In such cases the meaning is possessed 
by the mind both in the sense of this consistently 
determined attitude and in the further sense that 


Jiow this meaning should become explicit and what 
would be recognized as essential, when the attitude 
became self-conscious, is already implicit in the 
attitude itself* There is such a thing as confusion 
of attitude, reflected in hesitation of behavior and 
self-frustration, just as there may be inconsis- 
tency in the use of terms and in our explicit con- 
cepts. It is such hesitation and dubiety which 
provides the spur to that self -consciousness and 
self-criticism which renders meanings explicit. If 
meanings could not be present and determined in 
the attitude and behavior itself, there would be 
nothing to become conscious of. Objects do not 
classify themselves and come into experience with 
their tickets on them. The classifying attitude or 
mode of behavior which the mind brings to the 
given experience and which represents its mean- 
ing, dictates the explicit concept and implicitly 
possesses it already. 

If, however, in the light of this, it should be 
charged that I have used the phrase "common to 
two minds' 5 in a figurative and Pickwickian sense 
in the definition of "the concept," I shall plead 
guilty. I shall urge in extenuation that to begin 
the discussion by introducing all the qualifica- 
tion and explanation required for strict accuracy 
would have been cpnf using and impossible. I have 
but followed the custom to attributing to mind 
what is ideally determined by conscious attitude 
even though it is not explicitly present in con- 


sciousness. Whatever is merely convenient fiction 
in this can now be withdrawn in favor of an ac- 
curate equivalent. The concept is a definitive struc- 
ture of meanings, which is what would verify com- 
pletely the coincidence of two rm-nrfs when they 
understand each other by the use of language. 
Such ideal community requires coincidence of a 
pattern of interrelated connotations, projected by 
and necessary to cooperative, purposeful behavior. 
It does <not require coincidence of imagery or sen- 
sory apprehension. The concept is, thus, psycho- 
logically both an abstraction and an ideality, 
though in no greater degree or different sense than 
are most of things which are commonly attributed 
to mind. Both community of meaning and genuine 
understanding of reality are projected ideals more 
truly than realized actualities. \Ve study them as 
what our purposes intend and as that the approxi- 
mation to which gives value to our practice. It 
is concepts, as precisely such ideal abstractions, 
which must be implicity present in our practice, 
which constitute the element of interpretation 
which underlies our common understanding of our 
common world. To that topic, -we may now pro- 




The significance of conception is for knowledge. 
The significance of all knowledge is for possible 
action* And the significance of common conception 
is for community of action. Congruity of behavior 
is the ultimate practical test of common under- 
taking. Speech is only that part of behavior which 
is most significant of meanings and most useful 
for securing human cooperation. 

Common meaning may override all idiosyncrasy 
of feeling or sense, so far as such idiosyncrasy 
does not prevent congruous distinction and rela- 
tions. It may even override differences which are 
reflected in failure of discrimination and relation, 
in ways which have already been commented on. 
In fact, I think we may fairly be impressed with 
the tremendous achievement which our common 
meanings and our intellectual cooperation repre- 
sent in the life of the race. Community of meaning 
may also override much idiosyncrasy of behavior. 

*This chapter consists mainly of an elaboration and defense of con- 
ceptions put forward in the preceding one. It may be omitted, by any 
reader who chooses to do so, without prejudice to the understanding 
of later chapters. 



But eventually the very purposes for which com- 
munication exists insure a certain congruity of 
behavior when meanings are the same. 

Berkeley pointed out that we can never test the 
validity of knowledge by comparing an idea in 
the mind with an object outside the mind. We can 
only compare ideas among themselves. This is a 
pertinent consideration about the criteria of 
knowledge, whether one agrees with Berkeley's 
idealism or not. What I would here point out 
about the concept has a certain similarity. We 
cannot test community of meaning, even eventu- 
ally, by comparing the immediate experience in 
our own mind with the immediate experience in 
another mind, nor by comparing another's con- 
cepts, conveyed to us, with his immediate feelings 
and sensations. We can only compare meanings 
among themselves, as purely conceptual and ab- 
stracted from the character of any experience be- 
yond our own. We can only grasp another's mean- 
ings by observing the relation of his meanings to 
one another and to his behavior. 

In the end it can hardly fail to be the case that 
the possibility of our having concepts in common 
is conditioned by two things ; first, by the fact that 
we are creatures fundamentally alike, having in 
the large the same needs and interests and powers 
of discrimination and relation; and second, that 
we are confronted by a common reality, mediated 
to us in sense-experience which is comparable. 


Seeing that, in the large, such conclusions are 
indicated, there is a tendency to jump to them at 
the start on the part of rationalists to assume 
an ideal and complete agreement in an iron-clad 
and immutable set of categories, hypostatizing 
these as "human reason"; on the part of empiri- 
cists to presume that our common world is ex- 
hibited to all of us (that is, to all "normal" per- 
sons, the others being simply left out of account) 
in a common sensory experience. Theories of both 
these types are based on nothing more nor less than 
a beautiful myth. 

With the topic of the categories we are not just 
now concerned except for one point. Coincidence 
of categories may be interpreted to mean a neces- 
sary psychological coincidence of a certain aspect 
of experience, the formal or relational. Now I 
feel quite sure that something of this sort is true. 
My conviction is that one kind of likeness which 
is essential to our common understanding consists 
in certain very fundamental tendencies to action, 
growing out of basic similarity of needs and of 
physical structure ; these tendencies to action will, 
of course, be reflected in some aspect of experi- 
ence. But it is of some importance to see that even 
such fundamental tendencies in which we coincide 
need not necessarily be mediated by any direct 
psychological identity of experience. There seems 
to be the same possibility of systematic difference 
here that there is among our intuitions of length 


or of weight if one of us is tall and strong and the 
other short and weak. I should not stress this 
otherwise trivial point if it were not that a theory 
which, admitting divergence of experience at the 
level of ordinary perception, still bases itself upon 
a psychological identity of an esoteric and rec- 
ondite "human reason, 55 seems to me just the re- 
verse of what a sensible account of the common 
aspects of our common experience should be. Psy- 
chologically "human reason" is a very remote ab- 
straction; if the conception is to be retained, its 
aura of the transcendental needs to be removed. 

As for that other presumption, that our com- 
mon understanding is based upon the presence to 
us of a common world, it too is unduly simple. Our 
common world is very largely a social achievement 
an achievement in which we triumph over a 
good deal of diversity in sense-experience. Com- 
mon understanding would become progressively 
more difficult as community in what is given in ex- 
perience should become more meager; so much, 
the example of intelligent persons with defects of 
sense makes evident. But if we inquire at what 
point the limit may be set, or just what items of 
sense are absolutely requisite, we shall see, I think, 
that there is nothing of which we can say abso- 
lutely, "Unless this much were common in what is 
given to us, we could not understand each other 
at all." 

For these reasons, the problem of the genesis 


of common concepts will bear a little further con- 

In the first place, it may be pointed out that 
an initial community between "likeminded" indi- 
viduals is capable of enormous expansion, and that 
the manner of such expansion is familiar. 

Suppose two men speaking different languages 
but having a few words in common to be chained 
to the opposite walls of a dark cell, so that the 
possibility of establishing common meanings by 
such methods as pointing and naming would be 
at a minimum. With good luck in the initial com- 
mon concepts, and with a high enough order of 
intelligence, they might eventually establish a very 
large range of common notions by methods which 
the reader can imagine for himself. 

The actual case of Miss Helen Keller, in which 
a normal range of understanding has been de- 
veloped from original coincidence in kinesthetic 
and contact-sensation {absolute coincidence even 
in these being somewhat doubtful) need only be 

Or suppose there should be creatures on Mars 
of a high order of mentality. They might be psy- 
chologically rather different from ourselves and 
have senses and experience largely incomparable 
with our own. Yet if we could establish some 
initial common understanding (say if we should 

signal to them in light-flashes ~, - -, , and they 

should eventually respond with , , 


) ? then in spite of our differences from 

them, it would be hard to set a limit beyond which 


it could be said with certainty that our common 
understanding could not go. 

In fact, it is just this indefinite extensibility of 
conceptual understanding which is exhibited by 
abstract mathematical systems. To take the best 
illustration ; by a miracle of patience and insight, 
Mr. Russell and Mr. Whitehead have achieved, in 
"Principia Mathematical such an analysis of va- 
rious branches that the whole field of this subject 
(excepting the geometrical, so far omitted) can 
be developed from seven initial concepts. These 
undefined ideas are of such a sort that they must 
almost inevitably belong to any creature which 
should be conscious of its own ways of acting and 
should possess the habit of communication ideas 
such as "proposition" and "either-or." Suppos- 
ing these notions to be common to two minds in 
any terms you please these two minds could, if 
their patience and intelligence should be sufficient, 
arrive eventually at a common understanding of 
the whole of mathematics. 

So far, I am only concerned to point out that 
an initial community could extend itself extraordi- 
narily. The exigencies of common life, the need 
of cooperation, the tendency to imitation in be- 
havior, and the enormously developed institution 
of human education using the term in its widest 
meaning all go to enforce just this sort of elab- 


oration and extension of any initial mutuality of 
human understanding. Idiosyncrasy is pretty sys- 
tematically suppressed. And what cannot be sup- 
pressed (abnormalities and deficiencies) we go 
about it most earnestly and ingeniously to get 
around. A relatively meager mutuality of con- 
cepts, given human powers of discrimination, ab- 
straction and relation, and our human social hab- 
its, would be sufficient as the initial foundation 
for our actual and most elaborate mutual under- 
standing. I am not trying to argue that ft is from 
such a meager basis of initial mutuality that the 
community of understanding actually develops, 
though obviously one could make out a pretty 
good case from the manner in which the infant 
acquires his social inheritance of ideas. I wish only 
to point out the fact that, given such a meager 
mutuality, elaborate common understanding could 
develop; and that to argue straight from our 
elaborate common understanding to an equally 
extended coincidence of felt qualities or given ex- 
perience, is unnecessary and fallacious. 

The same considerations are pertinent as against 
those who would hold that an initial community 
of categories, as a psychologically identical and 
miraculous endowment, is necessarily presumed. 
Any who hold this doctrine are likely to argue 
against the view here presented something as fol- 
lows :* 

*I am not imagining what such critics might say but reporting 
what, in substance, some of them have said. 


It is possible that you can escape postulating 
psychological identity of sensuous content at the 
first stage, in the analysis of meaning, by resolv- 
ing the content of the concept into relations, but 
you will be forced to come back to it in the sec- 
ond, or some later stage, because the apprehen- 
sion of relations must be common. It may be true 
that "red" or "an hour" is conceptually the same 
for A and B, not because felt red or felt duration 
is sensuously identical in the two minds but be- 
cause both define red as the first band of the 
spectrum and an hour as sixty minutes. But 
the critic continues relations themselves must be 
discoverable in experience in a manner not essen- 
tially different from that in which substantive 
terms are given. That X is bigger than Y, or is 
to the left of, or better than, or stands in any 
other relation to Y, is something which must be 
disclosed by some sort of felt, empirical quale of 
the X Y complex. Otherwise experience could 
never determine for us whether X is better than 
Y or stands in any other relation to !F; and ex- 
perience would in fact be irrelevant to the truth 
of judgments of relation which means, of course, 
all judgments. 

So far, the critic's point is well taken. As was 
indicated in Chapter II, relations in general are 
given in very much the same sense that other 
properties of objects are given. However, it is 
equally true that relations, as cognized, are the 


result of conceptual abstraction and of a setting 
in connection with what is not given, just as prop- 
erties in general are. That is, I should wish only 
to file an exception, or a caution, that just as the 
predication of let us say roundness is not 
merely a direct report of sense-content, so the 
predication of "greater than" or "predecessor of" 
involves something more than the given. Though 
relations in general are given in experience like 
other properties, and though the applicability to 
experience of the predication of any particular 
relation depends, for any individual, upon the 
possibility of that relation's being given to him 
in some experience, still exactly those relations 
which constitute the interpretation of what is 
given are such as are not, just now, given in the 
experience which is the subject of such conceptual 

The critic continues: If, however, the presence 
or absence of certain relations between X and Y 
is revealed to the mind by the presence or ab- 
sence of some identifiable characteristic of experi- 
ence, and if identical concepts in two minds means 
the coincidence of a certain pattern of relation- 
ships, then the assumption that two minds have a 
concept in common is an assumption of some iden- 
tity of the order of psychological content. Rela- 
tions are, of course, definable in general. So it is 
conceivable that if substantive concepts are ana- 
lyzable into relations, relations may be analyzable 


into relations of relations as we see in mathe- 
matics sometimes. But and now the critic reaches 
the point of his argument unless you are to have 
an infinite regress, somewhere you must come to 
an end in a psychological identity which is abso- 
lute, in what I should think of as the categories. 
These will have to be at once the underlying pat- 
tern of all human experience and the elemental 
structure of human reason. 

Now I am not specially anxious to controvert 
anything which such a critic has to urge. I should 
wish merely to make two points against him which 
must greatly qualify the force of his argument. 
If conceptual analysis discovers the meaning of 
substantive terms in definitive relations, and con- 
sonantly, the meaning of relations in relations of 
relations, it does not in the least follow that there 
is a regress here which must either be infinite or 
come to some absolute end-terms. And second, the 
kind of psychological identity which most plausi- 
bly belongs to common concepts of the basic sort 
meant by "categories" is not at all of the type 
most frequently contemplated by those who talk, 
in capitals, about "Human Reason." 

To take the last point first; whether we take 
our examples of "the categories' 5 supposed to rep- 
resent the structure of human reason, from the 
traditional historical sources, or take them with 
much better reason from those scientific analyses 
which reveal basic concepts of mathematics, phys- 


ics, and so on, in either case our categorial con- 
cepts, "substance and accident," "cause and ef- 
fect," "different from," "either or," are certain 
to be such as are exemplified in experience by a 
very wide range of heterogeneous sensory con- 
tent. If, for example, the category "different 
from" must be common to two minds before they 
could even begin to create a common understand- 
ing, then it must at once suggest itself that what 
empirically exemplifies this category will be itself 
most markedly divergent in particular cases. If 
we must seek some psychological identity which is 
the vehicle of the category "different from," then 
about the only plausible place to look for it will 
jbe in our own activity of distinguishing. Any 
'reasonably conceived set of categories will exhibit 
these as a very high order of abstractions, and 
abstractions most unlikely to be identifiable by 
any simple coincidence of empirical quale in what 
is brought under them. Psychologically they will 
reflect much more directly our ways of acting than 
they will the character of that upon which these 
acts are directed. 

Apart from a prepossession in favor of some 
transcendental and miraculous status for "the 
categories," I suspect that the point which such a 
critic seeks to make against the view here pre- 
sented, is that human experience, at the lowest 
level at which it can be discerned the level here 
indicated by "givenness" already possesses a 


structure which reflects the nature of "human 
reason. 5 ' This, I take it, is entirely erroneous. 
Though doubtless the general character of sen- 
sory content reflects the nature of the animal, it 
is not such differences which are attributable to 
human reason or categories; it is precisely such 
divergence which community of thought and con- 
cept may triumph over. If, for example, one ex- 
amines such a list of basic concepts as the primi- 
ti% r e ideas of "Principia Mathematical one may 
see, I think, that the reason I cannot teach my 
dog the calculus is not because empirical exempli- 
fications of these primitive concepts are not pos- 
sible, or even familiar, to him, but because he is 
not capable of making an abstraction which is 
not dictated directly by instinctive interest and 
because a structure of relations must be either 
very simple or strongly enforced by repetition 
without exception in order for him to hold it in 
his mind* There is no reason to think that the ab- 
sence of human categories affects the content of 
his given experience in the least. 

Other considerations, pertinent to this topic, 
will be set forth in Chapter VII and Till. But 
already it may be clear that the sense in which 
the categories "inform" experience is not any 
sense in which different sets of categories (for 
different kinds of creatures) would presume any 
corresponding difference in experience as given. 
As Royce was fond of insisting, the categories are 


our ways of acting. What we can distinguish as 
attributable to our own acts are not, and can not 
be, limitations in the content of the immediate 
experience which is acted on. 

If, then, what the critic means to urge is that 
we must have identical feelings of the relation of 
substance and accident or of "if then,' 5 etc., I 
can only say that I still do not see the neces- 
sity of this; that I regard the point as rather 
tenuous for argument; and that in any case I do 
not see its importance for the theory of concepts 
in general which is here presented. It has been 
admitted that within the individual mind every 
concept must have its correlation, directly or 
eventually, with specific sensory content. If it 
should be the case that this specific sense-content, 
in different minds, is, for all our basic concepts, 
Identical, it would still be true that this sense- 
content can never be conveyed. We are concerned 
with two things in our practical understanding 
of each other with communication and with be- 
havior. My concepts are, from the outside view 
of me which you have, revealed as modes of my 
behavior, including speech. My words must main- 
tain a certain relation to other words which I use 
and to the things I do. It is necessary that we 
should act alike, in fundamental and important 
ways, if we are to have a possible basis for under- 
standing one another. But it is not necessary that 
when we act alike we should feel alike, however 


The eventual aim of communication is the co- 
ordination of behavior; it is essential that we 
should have purposes in common. But I can un- 
derstand the purposes of another without pre- 
suming that he feels just as I do when he has 
them. The psychology of purpose is an especially 
difficult topic, but it would seem that what is most 
essential is a certain relation between anticipation 
and realization. If, in another mind, both what is 
anticipated and what is realized should be, in 
terms of immediacy, different than for me, I could 
still attribute this relation between the two to him 
when I observe him to behave as I do* In other 
words, I correctly attribute a certain purpose, like 
mine, to another if I observe that he performs an 
act like mine and suppose correctly that it is the 
result of an intention involving the same congru- 
ence between anticipation and result which I find 
in my own case. I do not need to suppose that 
either purposes in general or the content of this 
act in particular are, as items of immediate ex- 
perience, identical in his case and in mine, in 
order to "understand his purposes.* 5 

Let us return now to the second point which 
we have supposed to be urged by our critic that 
the notion of the concept as a relational pattern 
of meanings among themselves must eventually 
break down because the analysis of substantive 
terms into relations presumes a similar analysis of 
relational terms into relations of relations, and so 
on ; that hence we shall be confronted either with 


an infinite regress of a hopeless sort or with the 
necessity that there should be some end-terms of 
such analysis which are absolute and have their 
meaning exclusively in some different fashion, 
such as imagery, which must therefore be identi- 
cal in two minds which understand each other. 

One might indulge in a great deal of loose talk 
about such an issue without reaching any real 
clarity. But as it happens there are excellent ex- 
amples to which the discussion can be tied, and 
these examples show that our critic is entirely 
wrong* I refer to the systems of pure mathematics, 
in their modern form. Such a system is generated 
deductively from certain primitive ideas which are 
taken for granted* All other concepts in the sys- 
tem are defined in terms of these. Unless these 
primitive ideas possess meaning it may be said 
the whole system would be meaningless; they 
are the end-terms of that particular analysis of 
the field in question which the system represents. 
Now let us imagine for a moment that this branch 
of mathematics should be a closed field; that if, 
for example, it is arithmetic which we are consid- 
ering, then no concept which occurs in arithmetic 
has any meaning outside arithmetic or other than 
an arithmetical application. This is not, of course, 
true; and it is for this reason that the ordinary 
"linear 55 method of developing mathematical sys- 
tems can work so well. These initial notions really 
would be and usually are clear before one un- 


derstands just what the system is to develop, be- 
cause these concepts have an application and a 
meaning outside arithmetic. But suppose this were 
not the case; would it then be true that all the 
concepts of the system must remain forever mean- 
ingless? I think we can see that this would not 
be so, 

In the first place, we may remark that there 
are entirely different sets of undefined concepts 
which would serve equally well as a basis. It is 
quite generally true that the same deductive sys- 
tem can be developed in a number of different 
ways, the only limit to this number of alternative 
developments being the practical difficulty of find- 
ing alternative sets of initial notions which are 
sufficiently clear and are economical as to the 
number of required postulates. If we take two 
different sets of undefined ideas (and of corre- 
spondingly different assumptions in terms of 
them) from which the same system may be devel- 
oped, then all the undefined ideas of the one set 
may be defined in the other; they are defined in 
terms of ideas which, in the other case, they them- 
selves serve to define. In other words, we can enter 
into the complex network of mathematical mean- 
ings in a number of different ways, from different 
points of departure. 

Since this is so, it is quite obvious that there is 
no inherent simplicity in either set of undefined 
ideas, and that the comparison of terms to points 


in a spatial pattern, and of definitions to the 
tracing out of such patterns of relationship, is 
much more apt than that other metaphor which 
represents logical analysis as physical dissection. 
If the undefined idea is "simple" and the predicate 
which defines it "a complex," then what is to be 
taken as simple and what as complex, is merely 
a matter of convenience and in no wise a logical 

Furthermore, some of the undefined notions in 
any deductive development of a mathematical sys- 
tem are pretty sure to be relations or "opera- 
tions." There is the same possibility of choice here 
that there is for the "substantive" notions. That 
is to say, relations are not necessarily defined by 
relations of relations but in a manner essentially 
the same as substantive terms. The supposition on 
which our critic charges our conception with re- 
quiring either an infinite regress or some absolute 
end-terms, is a natural one but quite erroneous. 
It is based upon the false analogy of logical to 
physical analysis. To discuss the logic of relations 
and of substantives so as to disclose general prin- 
ciples in precise language, would take us too far 
afield. Briefly, we may observe that it is as easy 
to define relations by the terms they connect as 
to define substantive terms by other substantives 
related to them. There is a certain analogy here 
to the fact that lines may be defined by the points 
they connect or points defined by the intersection 


If it should be said that this range of choice 
amongst the concepts of mathematics is due to the 
fact that all mathematical concepts are complex, 
and that every mathematical system presumes no- 
tions of a more fundamental sort which are used 
in it without explicit mention, then we may point 
out that in "Principia Mathematical 3 where aU 
the concepts of mathematics are defined or ana- 
lyzed, they are generated from the basis of con- 
cepts of logic ; and that the deductive development 
of logic itself presents exactly the same picture 
as here outlined. There is here the same range of 
choice as to undefined notions and postulates. In 
fact, the most economical development of logic 
yet discovered requiring only three symbolic 
postulates is in terms of an undefined idea so 
unob\ious that most people misinterpret it until 
they see precisely how it is related to other notions 
in the development based upon it. 

Now mathematical systems are by far the most 
extended and exact examples of logical analysis 
that we have. All such examples illustrate the 
fact that there is no such tiling as intrinsic sim- 
plicity or indefinability. All meaning is relational. 
Deductive order is, to a considerable extent, a 
matter of choice and is, in fact, usually determined 
by practical considerations of economy of assump- 
tion and the like. The mathematician does not 
choose his undefined ideas with any thought that 
they must be better understood than those which 


he intends to define, any more than he chooses his 
initial postulates in terms of them on the ground 
that these assumptions will be readily agreed to 
while the theorems he intends to prove might not. 
Often, in fact, the mathematician does not tell us 
explicifclv what things he is talking about but 
assumes "A class K of terms a, 6, c, * . . and a 
relation R or an operation such that . . , ," and 
leaves it to the postulated relations and the de- 
velopment of the system itself to identify his 
meanings* In general the theorems are no more 
"proved true 55 by the postulates than the postu- 
lates are by the fact that they lead to the theo- 
rems, and the terms defined are no more made 
clear by their definitions than the undefined terms 
are by the definitions into which they enter. What 
is, in fact, essentially demonstrated in a deductive 
system is the total fact of the order and connected- 
ness which is exhibited by the system as a whole. 
Even if the undefined ideas do happen to be clear 
initially, while those defined are not, nevertheless 
the development of the system serves to enrich and 
explicate those original meanings. The signifi- 
cance of the original notions is made clear by the 
relationships into which they enter, much as the 
significance of an hypothesis is increasingly ob- 
vious in any considerable survey of its conse- 

To bring this discussion back to its connection 
with the earlier point: If the mathematical sys- 


tern were a closed field, then the originally unde- 
fined concepts would of necessity possess the whole 
of their meaning in the extended order and inter- 
relationships of the system itself* The concepts of 
a particular mathematical system do not, of 
course, represent such a closed field. But the field 
of our concepts altogether is, and must be, closed 
in this sense. It is this fact which has been re- 
ferred to as the inevitable eventual circularity of 
definition and illustrated by the example of the 
ideal dictionary. Relations, we may further note, 
would be defined in such a dictionary as well as 
substantives. The conceptual meaning of terms is 
to be found in the array of their definitive or de- 
ductive relations to one another. That the rela- 
tions themselves are definable, does not lead to any 

Very likely it will be urged that, so far, I have 
ignored the large and important part which is 
played in the identification of meanings by the 
common reality which is presented. I have so far 
ignored it for the reason obvious to the reader, I 
hope that this "common reality 55 is precisely one 
of the things which needs to be accounted for, in 
the face of the fact that we cannot reasonably 
suppose that presented or immediate experience is 
actually common to the degree that reality is. 

Meanings are identified by the relational pat- 
terns which speech and behavior in general are 
capable of conveying. The sensuous content of 


experience in one mind cannot be conveyed to an- 
other, but the characteristic order of some set of 
items in the experience of A can be identified by 
B as belonging exclusively to some set of things in 
his own experience. The presence of like interests, 
if such may be presumed, will narrow the field 
within which search for such conceptual identity 
of order is to be made, and assist identification. 
Most such identifications of meaning will, of 
course, be based upon previous identification of 
other and related meanings. But the higher the 
order of intelligence, the greater the capacity to 
identify concepts simply and directly by their 
logical structure. A mathematician, for example, 
confronted by a system in entirely novel notation 
or in a language strange to him, might identify 
it in just this way. The complex numbers he 
might say are the only things in mathematics 
which have just that type of order. It is thus that 
we may imagine that intelligent Martians might 
catch our meaning in sending successively larger 
numbers of light-flashes. It is in such fashion 
that the meaning concealed in a cipher is finally 
disclosed by finding that rule which turns it into 
something which makes sense. 

In general, we are able to understand one an- 
other because for one reason a common real- 
ity is presented to us. But so to put it is to re- 
verse the order of knowledge. We have a common 
reality because or in so far as we are able to 


identify, each in his own experience, those sys- 
tems of orderly relation indicated by behavior, and 
particularly by that part of behavior which serves 
the ends of cooperation. What this primarily re- 
quires is that, in general, we be able to discrimi- 
nate and relate as others do, when confronted by 
the same situation. 

Although different individuals may, and to a 
certain extent verifiably do, intuite tilings differ- 
ently, still the basic discriminations which one can 
make can also be made by another. Especially 
those distinctions and relations which concern our 
major purposes and hence are such as it is prac- 
tically most important for us to discern in our ad- 
justment of behavior to environment, will be made 
by different individuals in comparable ways. Or 
to put it the other way about, we are "like crea- 
tures" and capable of understanding one another 
if, regardless of the sense-quality of what we in- 
tuite, we make the major discriminations and rela- 
tions concerned by the adjustment of behavior to 
environment in comparable ways. 

The "common reality" projected by such un- 
derstanding of each other is, to an extent not usu- 
ally remarked, a social achievement. It triumphs 
over a good deal of verifiable difference in the 
power of individuals to discriminate and relate in 
the presence of the same situation. The need to 
cooperate is always there. This being so, the im- 
portance of those concepts which are framed in 


terms of distinctions and relations which are com- 
mon, is enhanced, and of those which should be in 
terms of what some only can discriminate, is di- 
minished. If these distinctions which only some 
can make directly in the content of their experi- 
ence, do not concern what is important for behavior 
adjustment, then very likely no socially current 
concept will be framed in terms of them. There 
will be no language to describe these personal and 
peculiar phases of experience. And remembering 
how largely our thought is informed by social re- 
lationships it is likely that these phases of ex- 
perience will largely pass unnoticed by the indi- 
vidual himself. Again, even if they are noted, their 
significance will be regarded as "subjective" 
rather than of objective reality.* 

However, if, or in so far as, those distinctions 
and relations which can be made by some only are 

*In the end, the supposition of a difference in immediate experi- 
ence which is not to be detected through divergence in discrimination 
and relation, is a notion very difficult to handle. Because such dif- 
ference would, ex hypothesi 9 be ineffable. We can have no language 
for discussing what no language or behavior could discriminate. And 
a difference which no language or behavior could convey is, for pur- 
poses of communication., as good as non-existent. But this considera- 
tion only serves to enforce the fact that the assumption of qualita- 
tively identical immediate experience is unnecessary for community 
of knowledge that it is germane at all only so far as it affects that 
pattern of relationships here called the concept. 

The only reason that the possibility of such ineffable individual 
difference of immediacy is not altogether meaningless, is that we have 
interests which pass beyond those of cognition. Interests such as 
those of appreciation, sympathy, love, concern the absolute identity 
and quality as immediate of other experience than our own. Esthetics, 
ethics, and religion are concerned with such interests, which tran- 
scend those of action and of knowledge, as that term has here been 


important for survival or for the behavior-ad- 
justment required for the satisfaction of impor- 
tant needs, then like the blind man following his 
companion who sees we shall interpret "reality" 
in terms of the more differentiated experience of 
the better discriminator. Others will attach to con- 
cepts so framed some indirect meaning in terms 
of other aspects of their own experience; if in no 
other way, then in terms of the observed behavior 
of other persons. 

That we like-minded creatures have presented 
to us a common reality might seem to be, like the 
preestablished harmony of Leibnitz, simply a ma- 
jor miracle which must be accepted as a fact, 
whether we forthwith hypostatize that fact as "in- 
dependent reality" or not. But this miracle is in 
some part only the result of looking at the situa- 
tion wrong way to. We do not expect to have a 
common reality with an insect or an imbecile. 
"Like-mindedness" consists primarily of three 
things; the possession of like needs and of like 
modes of behavior in satisfying them, second, the 
possession of common concepts, represented in be- 
havior by discrimination and relation, and third, 
the capacity (evoked particularly when commu- 
nity in the other two respects threatens to fail) 
of transcending our individual limitations of dis- 
crimination by indirect methods. This last is a 
considerable item in what is meant by "intelli- 
gence." In short, the power to attain, directly or 


indirectly, to common concepts, applicable in com- 
mon ways, is itself the criterion of like-minded- 
ness. Such like-mindedness requires either a con- 
siderable community of order directly identifiable 
in experience or a considerable degree of intelli- 
gence by which disparity in the first respect may 
be compensated for. A Martian might be like- 
minded with ourselves in spite of quite different 
immediate experience. But, if so, he must be very 
intelligent. And such like-mindedness is pre- 
requisite to having a reality in common. 

When we remember that amongst "normal" in- 
dividuals there is very considerable variation in 
the acuity of perception, revealed in individual 
differences of discrimination, and when we remem- 
ber how plausible it is that other individual dif- 
ferences in presented experience exist but escape 
notice because of their small importance to our 
major needs, or because we learn our conceptual 
interpretations largely through imitation and co- 
operation, it becomes evident that the significance 
of the above considerations is by no means con- 
fined to the situation as between "normal" and 
"def ective" perceivers. 

The eliciting of "reality" from that presented 
experience in which the subjective and the ob- 
jective are jumbled up together, is an achieve- 
ment of intelligence expressed in our categorial 
distinctions. Our common reality reflects our com- 
mon categories. But it is both unnecessary and 


implausible to assume this fundamental commu- 
nity to be simply ready-made and miraculous. It 
seems much more reasonable to allow that this 
major outline of common reality reflects, in some 
degree, our common needs, our social organization 
for fulfilling them, and our learning from social 
example. Thus even our common categories may 
be, in part, a social achievement of like-minded- 
ness. The sharing of a common "reality" is, in 
some part, the aim and the result of social cooper- 
ation, not an initial social datum, prerequisite to 
common knowledge. 

To sum up, then: The purely conceptual ele- 
ment in knowledge is, psychologically, an abstrac- 
tion. It is a pattern of relation which, in the in- 
dividual mind, is conjoined with some definite 
complex of sense qualia which is the referent or 
denotation of this concept and the clue to its ap- 
plication in presented experience. These two to- 
gether, the concept and its sensory correlate, 
constitute some total meaning or idea for the indi- 
vidual mind. As between different minds, the 
assumption that a concept which is common is cor- 
related with sensory contents which are qualita- 
tively identical, is to an extent verifiably false, is 
implausible to a further extent, and in the nature 
of the case can never be verified as holding even 
when it may reasonably be presumed. Neverthe- 
less, community of meaning is secured if each dis- 
cover, within his own experience, that complex of 


content which this common concept will fit. When 
the behavior of each, guided by this common con- 
cept, is comparable or congruous, we have, so far, 
a reality in common. The traditional argument of 
the skeptic, that knowledge or the communication 
of ideas is dubious or impossible in the light of 
the subjectivity of sense, is without valid foun- 
dation. That our possession of any considerable 
array of common concepts depends upon the pres- 
ence to our minds of a common reality is or 
should be a commonplace. But both our com- 
mon concepts and our common reality are in part 
a social achievement, directed by the community 
of needs and interests and fostered in the interest 
of cooperation. Even our categories may be, to a 
degree, such social products; and so far as the 
dichotomy of subjective and objective is governed 
by consideration of community, reality itself re- 
flects criteria which are social in their nature. 


We have so far taken the clue of common 
meaning of what can be conveyed from one 
mind to another as the criterion of the concept. 
But, as has been noted, this was in part merely 
an expository device. However much our concepts 
are shaped by social intercourse and borrowed 
ready-made by the individual, a human being 
without fellows (if such can be imagined) "would 
still frame concepts In terms of the relation be- 
tween his own behavior and his environment. 
Knowledge must always concern principally the 
relations which obtain between one experience and 
another, particularly those relations into which 
the knower himself may enter as an active factor. 
It is the given as thus conceptually interpreted 
which is envisaged as the real object. 

It is also true that exclusive emphasis upon the 
social, or the talcing of language as a point of 
departure, might easily lead to an oversimplifica- 
tion of our notions of conceptual interpretation. 
TVords represent rather large and ready-made 
wholes relatively stable and relatively simple 
concepts which are a somewhat loose fit for the 
precise and complex knowledge of perceived ob- 



jects. In a glance of the eye, so to speak, we ap- 
prehend what whole paragraphs will do no more 
than suggest. Language is primarily useful for 
conveying generalizations or else very specific ab- 
stracted items of experience. Not only is that 
knowledge of an object which is mediated by per- 
ception something which is usually difficult to con- 
vey precisely in words but usually it is not im- 
portant to convey it in more than very partial 
fashion, since those who are required to act di- 
rectly toward what is presented to us are usually 
those who are also present to it themselves. 

In fact, this difference between what words con- 
vey and what perception mediates is so marlfed 
that it may suggest a distinction of two kinds of 
knowledge; direct knowledge of objects (ac- 
quaintance with), gained by the presentation of 
them in experience and immediately verifiable, and 
propositional knowledge or generalization (knowl- 
edge about) which concerns more than can be 
given at one time and thus requires some mental 
synthesis of what is temporally disjoined* 

Such a dichotomy, however, would be falsely 
taken. It is the first thesis of this chapter that 
there is no knowledge merely by acquaintance; 
that knowledge always transcends the immedi- 
ately given. The merely contemplated or enjoyed 
may possess esthetic significance, but if it is to have 
cognitive meaning this immediacy must become 
the subject of an interpretation which transcends 


it; we must take toward the given some attitude 
which serves practical action and relates it to what 
is not given. Let us first briefly illustrate the na- 
ture of such conceptual interpretation. We may 
then turn to the special problems which are in- 

At the moment, a certain "that" which I can 
only describe (in terms of concepts) as a round, 
ruddy, tangy-smelling somewhat, means to me 
"edible apple. 55 Now my ultimate purpose toward 
it may be the enjoyment of an ineffable taste. But 
that taste not being given, I need a conceptual go- 
cart to get me over the interval between this 
round, ruddy presentation and the end projected 
by my purpose. Life is full of such undesirable 
interstices; in fact just so far as it needs to be 
earnest and active, it is made up of them. It is the 
function of mind to bridge these by assigning to 
the present given an interpretation through which 
it becomes related to, or a sign of, a correlation 
between certain behavior of my own and the real- 
ization of my purpose. This interpretation has 
the character of a generalization which has been 
learned. I phrase it by saying "That (denoting 
the given presentation) is a sweet apple (con- 
noting among other things the possible taste)** 5 
If I should be completely absorbed in the first 
given, as an infant might, then I should frame no 
concept, it would have no meaning, and no action, 
unless a merely instinctive one, would be evoked. 


An object such as an apple Is never given; be- 
tween the real apple in all its complexity and this 
fragmentary presentation, lies that interval which 
only interpretation can bridge. The "objectivity" 
of this experience means the *verifiability of a fur- 
ther possible experience which is attributed by this 

The notion that there is a simple sort of knowl- 
edge, gained by direct apprehension alone, has 
two major sources. In the first place it is falsely 
supposed that there are some concepts at least 
which denote "simple qualities" something which 
can be directly exhibited in a single experience. 
And second, the word "knowledge" is sometimes 
used for that enjoyment or contemplation which 
projects no purposes but is completely absorbed 
in the given as an esthetic object. (Whether there 
are any experiences which have exclusively this 
character is open to doubt, but at least experi- 
ence may have this ingredient or this aspect.) 
Putting these two together, it is easy to arrive at 
the erroneous conclusion that there is a kind of 
cognitive apprehension of simple qualities or es- 
sences which terminates directly in the given ; it 
may even be supposed that other knowledge rises 
out of this by some kind of complication and thus 
that direct awareness is the simplest and the basic 
type of knowledge. 

That there is direct apprehension of the imme- 
diate, it would be absurd to deny ; but confusion is 


likely to arise if we call it "knowledge." There are 
no "simple qualities 5 ' which are named by any 
name ; there is no concept the denotation of which 
does not extend beyond the immediately given, and 
beyond what could be immediately given. And 
without concepts, there is no knowledge. 

There are recognizable qualitative characters 
of the given, which may be repeated in different 
experiences, and are thus a sort of universals; I 
call these "qualia." But although such qualia are 
universals, in the sense of being recognized from 
one to another experience, they must be distin- 
guished from the properties of objects. Confusion 
of these two is characteristic of many historical 
conceptions, as well as of current essence-theories. 
The quale is directly intuited, given, and is not 
the subject of any possible error because it is 
purely subjective. The property of an object is 
objective ; the ascription of it is a judgment which 
may be mistaken; and what the predication of it 
asserts is something which transcends what could 
be given in any single experience. 

Consider such a property as "round" or "blue." 
The real roundness of the real penny is seen as all 
degrees of elliptical appearance; the blueness of 
the blotter may be seen as any one of a whole 
range of color-qualia, depending on the illumina- 
tion. The judgment that the penny is round may 
be made because it "looks round" or it may be 
made because, under given conditions which are 


understood, it "looks elliptical." But the given- 
ness of the appearance is not the givenness of ob- 
jective roundness. Given the elliptical appearance, 
the judgment "That is round/ 5 may be in error. 
Indeed, given the round appearance, the judg- 
ment may still be in error, as measurement with 
precision instruments might reveal. A thing which 
looks blue under a certain light may be blue or it 
may be green, and a thing that looks green or 
purple may be blue. A penny run over by a rail- 
road train will look round when held at a certain 
angle, while one which looks elliptical may be 
round. In other words, the same quale may be, for 
a correct interpretation, the sign of different ob- 
jective properties and different qualia may be the 
sign of the same objective property. 

The confusion of the quale and the objective 
property has doubtless come about through a 
short-cut in the use of language which is charac- 
teristic of common-sense. A thing is said to *%ok 
round" when it presents the quale which a really 
round object does when held at right angles to the 
line of vision; and a thing is said to "look blue" 
when it looks the way a really blue thing does un- 
der usual or standard illumination. In general, 
the name of the property is also assigned to the 
appearance of it under certain optimum condi- 
tions. The round penny looks round when held 
at that angle at which judgment of actual shape 
from visual appearance is safest. And an object 


looks the color that it is under that illumination 
which is conducive to accurate color discrimina- 
tion. A thing looks as big as it is at about that 
distance (for objects of its size) at which human 
beings make fewest mistakes in judgments of 
magnitude. This use of language has its obvious 
practical motives, but it would be an extraordi- 
narily poor observer who should suppose that 
what the name means in ordinary parlance is the 
appearance as such and not the objective prop- 

It is not, of course, a philosophic problem to de- 
termine how such language should properly be 
used. But it is worth remarking that those philoso- 
phers who suppose that the names of properties 
are first the names of certain given qualia and 
therefore of the properties of objects which, un- 
der optimum conditions, present them, have 
missed something significant which determine the 
common-sense use of language. 

Qualia are universals, and they are universals 
such that without the recognition of them by the 
individual nothing presented in experience could 
be named or understood or known at all. At this 
point it would be very easy to fall into con- 
troversy about the use of language which above 
all things I wish to avoid. Whether one should say 
that there must be concepts of qualia because they 
are recognized, or no concepts of qualia because 
they are ineffable; whether the immediate appre- 


hension of qualia should be called "knowledge" 
because of its function in the cognition of objects, 
or should not be called "knowledge" because it 
neither needs nor can have any verification; 
whether this direct awareness should be merely so 
designated or should be termed a "judgment" all 
this has to do only with the meaning of the terms 
"concept," "knowledge," "judgment." What I 
wish to point out is the real and important dis- 
tinction between qualia and the immediate aware- 
ness of them on the one hand and the properties 
of objects and our knowledge of them on the other. 
Qualia are subjective; they have no names in 
ordinary discourse but are indicated by some cir- 
cumlocution such as "looks like"; they are in- 
effable, since they might be different in two minds 
with no possibility of discovering that fact and 
no necessary inconvenience to our knowledge of 
objects or their properties. All that can be done 
to designate a quale is, so to speak, to locate it in 
experience, that is, to designate the conditions of 
its recurrence or other relations of it. Such loca- 
tion does not touch the quale itself; if one such 
could be lifted out of the network of its relations, 
in the total experience of the individual, and re- 
placed by another, no social interest or interest of 
action would be affected by such substitution. 
What is essential for understanding and for com- 
munication is not the quale as such but that pat- 
tern of its stable relations in experience which is 


what is implicitly predicated when it is taken as 
the sign of an objective property. 

Apprehension of the presented quale, heing im- 
mediate, stands in no need of verification ; it is im- 
possible to be mistaken about it. Awareness of it 
is not judgment in any sense in which judgment 
may be verified; it is not knowledge in any sense 
in which "knowledge 35 connotes the opposite of 
error. It may be said that the recognition of the 
quale is a judgment of the type, "This is the same 
ineffable 'yellow* that I saw yesterday." At the 
risk of being boresome, I must point out that 
there is room for subtle confusion in interpreting 
the meaning of such a statement. If what is meant 
by predicating sameness of the quale today and 
yesterday should be the immediate comparison of 
the given with a memory image, then certainly 
there is such comparison and it may be called 
"judgment" if one choose; all I would point out 
is that, like the awareness of a single presented 
quale, such comparison is immediate and indubita- 
ble; verification would have no meaning with re- 
spect to it. If anyone should suppose that such 
direct comparison is what is generally meant by 
judgments of qualitative identity between some- 
thing experienced yesterday and something pre- 
sented now, then obviously he would have a very 
poor notion of the complexity of memory as a 
means of knowledge. He might be advised to try 
buying a spool of thread to match something left 


at home. The usual statement, "This is the same 
yellow I saw yesterday, 35 truly represents a judg- 
ment because at least one of the things compared 
is an objective reality a temporally continuing 
entity which retains its identity and character. 
This meaning is something which could be veri- 
fied, under conditions which are conducive to the 
permanence of color, by going back to the object 
seen yesterday, or in some other, and perhaps in- 
direct, fashion. The judgment is about an ob- 
jective property of a thing. To suppose that a 
quale itself is such an enduring entity is to work 
confusion between what is immediate and some- 
thing else which, from the point of view of knowl- 
edge, is an intellectual construction of a highly 
complex sort.* 

An immediate quale apart from some relational 
context which "locates" it in experience is intrin- 

J*Immediate comparisons are presumably very important in deter- 
minations of value. Such determinations of better and worse, as be- 
tween immediately presented qualia such as two pleasures or pains, 
conjointly experienced, are often called "value-judgments.** Such 
comparison is doubtless indispensable to determination of value, 
but, here as elsewhere, judgment, in the ordinary sense, does not 
concern the immediately presented as such or for its own sake, but 
the enduring value of something objective to which this immediate 
comparison may be a clue. Here as elsewhere, the immediate com- 
parison is indubitable and verification has no meaning for it; what 
needs to be verified and is worth judging is the permanent quality of 
some object, or type of objects, or some permanent possibilitv of 
value-experience. * 

It may be further noted that it is this confusion of subjective and 
objective, referred to in the text above, which almost inevitably re- 
^ te th mte Pokt>? of a "sensum" between the subject and 
the object. If a sensum has character which endures and can be 

3fm nM^ ?J^ F " not noticed '." the * a ^rge part of the 
problem of knowledge concerns our veridical apprehension of sensa; 
the supposition that they simply are in mind or identical with the 


sically and absolutely inarticulate. It is inarticu- 
late not only in the sense that it cannot be ex- 
pressed to another; it would be impossible for it 
to be abstracted and envisaged as an object of our 
own thought. Imagine a man to suffer all his life 
from toothache, but in such wise that no pres- 
sure on the jaw, no change of temperature or of 
the heart-beat, no behavior of himself or difference 
of surroundings would ever alter it. Not only 
would a person so afflicted be unaware that he suf- 
fered toothache it would not in fact be a tooth- 
ache; he could not even become conscious of the 
ache as a distinct fact of his experience* He would 
be aware of it as the cow is aware of hunger, per- 
haps, but it would never become for him an ex- 
plicit object of thought. Such an all-pervasive 

content of awareness is incompatible with the possibility of erroneous 
judgment of their enduring and objective character. 

Also, it may be said that the statement St This is the same yellow 
that I saw yesterday" has meaning in a sense which does not have to 
do with the objective quality of a physical thing. I believe this is cor- 
rect; it may intend the assertion of the qualitative identity of two 
psychological states. But the statement with this meaning is significant 
only in the same sense in which it is verifiable; that is, on the assump- 
tion that psychological states are events which modify a substantive 
thing, the mind, and that this enduring mind is an organized reality 
in which, as in physical things, events are later verifiable by the 
effects which they produce. TVhat is not "thing" or objective, in 
terms of our knowledge of the physical, may be something objective 
in the categories of psychology, Qualia as presentations of external 
reality and qualia as states of a mind are quite different matters. In 
both cases, they are presentations of objects quite different objects 
because the relational context into which the presentation is brought 
in being understood is quite different in the two cases. In the one 
case, they are presentations of an external reality, a physical thing 
or property; In the other, they are presentations of a psychic reality, 
a mind. In no case are they objective or the object of knowledge 
apart from a relational context of conceptual interpretation. And in 
all cases, the judgment or knowledge is about what is objective. 


ingredient of experience could never become ar- 
ticulate, because it would lack the ground of any 
possible discrimination and relation. No language 
could express it ; there would be no thinking to be 
done about it; there would be no possibility of 
bounding it or eliciting it as a separate fact. For 
the person in question, it would be a part of his 
life or a coloring of reality but no part of his 
knowledge. The concept of a toothache does not 
consist of the ache but broadly speaking in the 
apprehension of what brought it on and the 
formula for getting rid of it. AH that is intel- 
ligible about it is the set of relations in which it 

There are no concepts of immediate qualia as 
such not because the word "concept" as here 
used has any unnecessary connotation of the ver- 
bal, but because articulation is the setting of 
bounds and establishing of connections; because 
what does not affect discrimination and relation 
has no handle by which the mind can take hold 
of it. It would be erroneous to take this fact to 
mean some positive bafflement in the presence of 
the immediate, because there is here no question 
we can ask which fails to find an answer ; there is 
no interest which is baffled. The interest of knowl- 
edge is for action, and action proceeds by way of 

If, then, we take "simple ideas" such as "blue" 
and "round" as, so to speak, the least concepts 


that there are, we find that what such concepts 
embrace Is not an immediate quale as such but 
some stable pattern of relations. We have con- 
cepts of objective properties these are indicated 
by the manner in which we should proceed to 
verify the blueness of the blue blotter or the 
roundness of the penny. The manner in which such 
verification must take place is obvious. To verify 
a shape, we walk around the object or manipulate 
it. Its successive perspectives and their relation 
to our behavior meanwhile, must present a cer- 
tain order of temporal relationships. If it is a 
small object, we may corroborate the visual by 
tactile and kinesthetic impressions. If a large one, 
we may measure it. Feeling the roundness of a 
marble as we roll it between thumb and fingers, 
or measuring a house, is again a temporally ex- 
tended and ordered relation of apprehended 
qualia. To verify a color, we change the condi- 
tions of illumination or alter the angle so as to 
get rid of the sheen, or we bring the thing into 
juxtaposition with some object whose color has 
previously been tested or is accepted as a stan- 
dard of comparison. When we thus manipulate the 
object or behave toward it, we must know what we 
expect if it really is round or blue, or what not. 
If what is thus predicted does not supervene, then 
our first ascription of the objective property will 
be withdrawn as proven false* The objective real- 
ity of the property consists, of course, from the 


point of view of knowledge, of what would verify 
It, and includes everything the failure of which 
would lead to the withdrawal of the judgment as 
mistaken. The concept of the blueness or round- 
ness of an object which is presented includes all 
that is essential to the truth of the predication of 
the property. Thus what constitutes the existence 
of an objective property and the applicability of 
a concept even of the simplest sort is not a 
given quale alone but an ordered relation of dif- 
ferent qualia, relative to different conditions or 
behavior* This pattern or order, which is what the 
adjective names, will always be temporally ex- 
tended (which is the same as to say that the predi- 
cation of the property is something verifiable) , 
and always it will have relation to our own possible 
ways of acting toward the presented object. 

If then, one approach the problem of our 
knowledge of objects as many have done and do 
with the notion that the object is "known as" 
some complex of presented or presentable quali- 
ties, and is thus analyzable into "simple" quali- 
ties which are capable of being presented and 
identified in a momentary experience and are es- 
sences or the denotation of certain simple con- 
cepts, I hope the nature of his error will be clear. 
If concepts are to be articulate and meaningful, 
then the application of them must be something 
verifiable; which means that what they denote 
must have a temporal spread. Not a momentary 


presented quale but an ordered relationship of 
such, is the least that can be meaningfully named. 
The predication of a property on the basis of 
momentarily presented experience, is in the na- 
ture of an hypothesis, which predicts something 
definitely specifiable in further possible experi- 
ence, and something which such experience may 
corroborate or falsify. The identifiable character 
of presented qualia is necessary to the predication 
of objective properties and to the recognition of 
objects, but it is not sufficient for the verification 
of what such predication and recognition im- 
plicitly assert, both because what is thus asserted 
transcends the given and has the significance of 
the prediction of further possible experience, and 
because the same property may be validly predi- 
cated on the basis of different presented qualia, 
and different properties may be signalized by the 
same presented quale, If the denotation of any 
concept were an immediately apprehensible quale 
or complex of such, then the ascription of this 
concept when such qualia were presented could 
not conceivably be in error. That the predication 
of a property may be mistaken, or the perception 
of it illusory, corroborates the fact that what is 
involved in the cognition of the property tran- 
scends the given. One cannot be mistaken about 
the content of an immediate awareness. If I have 
bitten an apple, I cannot be mistaken about the 
taste in my mouth. But I may conceivably fall 


into error in predicating sweetness of the apple or 
expecting a similar taste from another bite. The 
only sense in which apprehension could be illusory 
or erroneous is just this sense of including a mean- 
ing which is not given but is added to the given 
and must be verified, if at all, in some other ex- 

It is evident that the considerations here 
brought forward with reference to the knowledge 
of objective properties will hold for the knowl- 
edge of objects in general* Knowledge always 
transcends the immediately given. It begins with 
the recognition of a qualitatively specific presen- 
tation, but even that mmimum of cognition which 
consists in naming is an interpretation which im- 
plicitly asserts certain relations between the given 
and further experience. The ascription of a sub- 
stantive or an adjective is the hypothesis of some 
sequence in possible experience or a multiplicity 
of such sequences. The verifiability of such is es- 
sential to the nature ascribed to the object in rec- 
ognition, or even to the acceptance of this experi- 
ence of it as presentation of the real. The cri- 
terion of the objectivity of what is presented is 
always such a relation to further experience. In 
the nature of the case, the difference between 
veridical perception and an experience which is 
genuinely illusory (really deceptive to the indi- 
vidual in question) is never to be discovered within 
what is strictly given in the presentation. When 


we distinguish one experience as illusory, another 
as presentation of the real, we can intend noth- 
ing even conceivably verifiable except that, start- 
ing from the given experience and proceeding in 
certain ways, we reach other experience which is 
predictable in the one case and not in the other, 
Thus "acquaintance with" tJie recognition of 
what is presented as a real object of a certain 
Jcindy has already the significance of prediction 
and asserts the same general type of temporal 
connection as our knowledge of Icew^ tlie "knot&l- 
edge about" which is stated in generalizations. 
This is merely to reiterate Berkeley's doctrine of 
the "idea" as a sign, with the added thought that 
what is contained within any one idea or presen- 
tation is never more than a fragment of the na- 
ture of the real object. The ascription of this 
objectivity to the presentation is the conceptual 
interpretation of what is presented. 

On the other hand, that kind of knowledge 
which may strike us as more truly conceptual 
must always if it be knowledge of reality come 
down to just such interconnectedness of experi- 
ence and must be verifiable in the pattern of pre- 
sented experiences. It is this which is affirmed in 
the dictum of Charles Peirce: "Consider what ef- 
fects that might conceivably have practical bear- 
ings you conceive the objects of your conception 
to have. Then, your conception of those effects is 
the whole of your conception of the object." "Ef- 


fects" which are verifiable can, in the end, mean 
nothing more than actual or possible presenta- 

The perceptual knowledge of an object is 
neither the coincidence of mind and object in the 
momentary experience nor any duplication of the 
object in the mind. A presentation bespeaks the 
activity of the mind, first, in that it means an ab- 
straction and a setting of bounds within the total 
field of the given. And second, the manner of this 
abstraction already reflects a classification which 
is the implicit prediction of further experience. 
This excision of the presentation is its recogni- 
tion as the appearance of a thing its classifica- 
tion with other such presentations (of the same 
kind of thing)* Such classification reflects a gen- 
eralization from experience which predicts cer- 
tain orderly and lawful connections of this presen- 
tation with further experience. In the absence of 
such implicit prediction, the presentation would 
be meaningless. Thus the classification of what is 
presented and the predicted relationship of it 
with further experience are one and the same 
thing* This implicit prediction is at once a gen- 
eral principle and our concept of the object. 

The concept, to be sure, is substantive or ad- 
jectival, not prepositional, but the application of 
the concept to the presentation is an interpreta- 
tion of the same fundamental sort that proposi- 
tions express. The validity of such application de- 


pends on the subsistence of those relations to other 
experience which are implicit in the concept. It is 
such propositions which we implicitly assert in 
naming and recognizing, and do not think to ex- 
press. They are the knowledge upon which all 
other knowledge of reality is built. 

We have knowledge of objects, then, not 
through any coincidence of mind and object in 
awareness but precisely so far as this is tran- 
scended. The merely contemplated or enjoyed 
may possess esthetic significance but it has no 
cognitive meaning. The dictum of the new real- 
ists, that mind and object coincide so far as the 
object is just now known by this mind and so far 
as the mind is just now a knowing of this ob- 
ject, is as wrong as possible. So far as mind and 
presentation coincide, the state of mind is not cog- 
nition and the presented object Is not known. 
Presentative theories of knowledge in general as 
well as representative theories commit a similar 

For the object presented to be real, there must 
be more to it than could be given in any single 
experience. The objectivity of the experience Im- 
plies this "more." And this cognitively signified 
"more," which Is the meaning of the presentation, 
must be verifiable. But what does it signify that 
there should be verifiably more to any object than 
is given in the single experience of it? It can mean 
nothing else than the possibility of other expert- 


ences, of a predictable sort, related to this experi- 
ence in predictable ways. Any other kind of 
"more" attached to the presentation would be un- 
verifiable. If some will contend that there can be 
any other kind of "more" to the object, they must 
tell us how the existence of the unverifiable is to be 

To be sure, that "more" which is the verifiable 
meaning of the presentation, and our concept of 
the object, is normally not verified in any but the 
most fragmentary fashion. When I interpret a 
certain round ruddy presentation as "sweet ap- 
ple/' I implicitly predict that if I should bite, 
then I should get a certain taste. But perhaps I 
am not hungry now, so I merely file away this 
possibility for future reference. The "if " clause of 
my prediction is allowed to remain contrary to 
fact. Precisely here, in this apparently common- 
place fact that the meanings ascribed in our rec- 
ognition of objects and predications of properties 
must be verifiable, yet are commonly not verified 
or only verified in part, is something of great im- 
portance both for understanding the relation of 
knowledge to our ways of acting and for the na- 
ture of objectivity or thinghood. 

A mind for which, whatever happened, nothing 
could be done about it, could possess no knowl- 
edge, either of generalizations or of objects. Not 
only would knowledge in such a mind serve no pur- 
pose, and thus be a genetic puzzle; it would not 


exist at all. For the merely receptive and passive 
mind, there could be no objects and no world. 

A thing we may say as a first approximation 
is a complex of properties or qualities, recog- 
nizable by some uniformity of appearance ; a gen- 
eralization or law designates some uniformity of 
behavior. But, as we have seen, appearance and 
behavior are not wholly separable. An appearance, 
if it be that of a recognizable and nameable tiling, 
and not merely some meaningless and indescrib- 
able cross-section of the flux of experience, must 
recur in predictable ways. It must be a subject 
about which we can generalize. My desk-chair is 
not a universal entity but a unique object ; in my 
experience, however, and in that of the household, 
it is a reliable uniformity. Every morning I find it 
there before my desk. It is this recurrence, and a 
hundred other trivial uniformities of its behavior 
as a part of my experience, which constitute it a 
thing. If we consider some even more stable object 
say the vase on the mantel which stands always 
in one place and exhibits no alteration whatever 
that can be detected then, there would be no be- 
havior which is ascribable to the object itself; it 
does not change. But it is still because of its re- 
currence, in a certain context, in my experience 
that I recognize this presentation as a thing. That 
change which constitutes the background within 
which it persists as an identity, is a charge which 
I ascribe to my own behavior, my comings and 


goings. But it Is still true that if this recurrence 
in experience did not conform to certain generali- 
zations about the process or eventuation of ex- 
perience, it would not be a thing but an illusion, 
like the quite different object which I saw there 
last night in my dream. 

There may seem to be a fundamental categorial 
distinction between what is thing, designatable by 
a substantive, and what is transition, relation, 
some connection of things which is designatable 
only by a proposition and, when predictable, is 
known as generalization or law. It is this dichot- 
omy which, by being misunderstood, leads to the 
false division of knowledge into immediate "ac- 
quaintance with 55 and "knowledge about.' 5 We 
might suppose that this dichotomy correlates with 
the division of experience into what James called 
"flights and perches." An appearance of a thing 
is a kind of arrest; relation or the propositions! 
signifies the flux. But this, I think we may see, is 
false. Experience may be as full of flux as a trip 
in a roller-coaster, but if it is a flight which is 
repeatable under conditions which can be speci- 
fied, it is the appearance of a thing. It is true that 
thinghood connotes some stability and persistence, 
while "law" designates some uniformity of change 
or process* But the true basis of this dichotomy 
is not the division into flights and perches in ex- 
perience; it is the division between that part of 
the faix of experience which I ascribe to myself 


and m,y own activity r , and that change which can- 
not be so predicated of myself and is objective. 

Both thing and the change described by law be- 
long to the objective world. The objective reality 
of things is that uniformity I reach by a reorder- 
ing of the actual process of experience so as to 
subtract, or integrate out, "my own activity" ; ob- 
jective change is that part of the process which 
remains after such subtraction* That a thing is 
always there in the objective world 3 does not mean 
that it is always there in experience ; it means that 
it is always there when I look for it in the right 
place or in the right way. It means that its recur- 
rence in experience in uniform ways relative to 
my own action is predictable. "Looking for it 5 * 
may, of course, mean a quite complex process and, 
in certain types of cases, may reach the object 
only indirectly; also the kind of "stability'* or 
^^persistence** to be expected varies with the type 
of object and may be highly complex. But at bot- 
tom, when complications and qualifications have 
been dealt with, thinghood means a stability or 
uniformity of appearance which can be recovered 
by certain actions of my own. 

This distinction between objective change and 
those transitions of experience which are not ob- 
jective because attributable to myself, was first 
pointed out by Kant. The order of positions of the 
boat going down the stream is objective because 
by nothing I can do is this order in experience to 


be reversed. But the sides of the house are simul- 
taneous and are merely partial presentations of 
one stable and unchanging thing because the order 
of their succession in experience is due to my own 
action. It is obvious that we here confront a fun- 
damental consideration of the utmost importance 
for the analysis of the categories, for the distinc- 
tion of self and external world and for the division 
of the flux of experience into time and space. Into 
these most difficult matters, we cannot go, but it 
may be in point to observe the basic character of 
the conception of "activity." 

To ascribe an objective quality to a thing means 
implicitly the prediction that if I act in certain 
ways, specifiable experience will eventuate: if I 
should bite this, it would taste sweet; if I should 
pinch it, it would feel moderately soft ; if I should 
eat it, it would digest and not poison me; if I 
should turn it over, I should perceive another 
rounded surface much like this ; if I should put it 
on the scales, they would register about three 
ounces. These and a hundred other such hypo- 
thetical propositions constitute my knowledge of 
the apple in my hand. These are the meaning which 
this presentation 7ias to me now, but it may be 
that neither now nor in the immediate future do 
I actually verify these possibilities* 

It is only because we are active beings that our 
world is bigger than the content of our actual ex- 
perience. For the active being, reality is as much 


bigger than the content of given experience as is 
measured by the totality of all that is related to 
what is presented by those propositions of the type 
"If I should . . * then . . ." which he takes to be 
true- All that "more" which belongs to the objects 
presented to him, over and above what is immedi- 
ately given, and all the rest of reality, as it stands 
related to his object but not presented with it, re- 
sides in this potency of possible experience. For 
the passive being, the only possible passage of ex- 
perience is the actual one : the only continuities of 
reality the only relations with any given would 
be the actual flux of experience. No object would 
be thicker than its presentation. And since all mo- 
tive to that analysis which delimits the presenta- 
tion would be lost in the lapsing of any manner of 
relating it except to the actually antecedent and 
consequent items of the given, the distinction of 
the given field into presentation and background 
would serve no purpose and would presumably fail 
to be made. Thus for the passive being the whole 
of reality would collapse into the actual proces- 
sion of the given ; indeed it must collapse into the 
specious present, since the objectivity of memory 
and anticipation is a complex interpretation put 
upon presented imagery, Even the "stream of 
consciousness" itself is a highly conceptual con- 
struction, requiring, amongst other items, the cat- 
egory of objective time as the order of that change 
which is irreversible by the "if" of any altered 


mode of action. For the passive being, there would 
be no distinction of subjective and objective ; the 
given would be both subject and object, and the 
whole of both. Nothing would be verifiable save 
what was presently verified, and hence nothing 
would mean anything. There could be no distinc- 
tion between real and illusory. Since nothing given 
would transcend its givenness and there would be 
no distinction of real from unreal, there would be 
no reality. 

The whole content of our knowledge of reality 
is the truth of such "If then" propositions, in 
which the hypothesis is something we conceive 
could be made true by our mode of acting and the 
consequent presents a content of experience which, 
though not actual now and perhaps not to become 
actual, is a possible experience connected with the 
present. For the active being such hypothetical 
propositions can be meaningful and true when the 
hypothesis is false.* The attribution to what is 
given of connection with such a content of further 
possible experience is the conceptual interpreta- 

*The significance of knowledge depends upon the significance of 
possibility that is not actual. Possibility and impossibility hence 
necessity and contingency, consistency and inconsistency, and various 

other fundamental notions require that there should be "If 

then " propositions whose truth or falsity is independent of the 
truth or falsity of the condition stated in their antecedent clauses. 
Those readers who happen to be familiar with my doctrine of "strict 
implication" will find here a motive for the distinction of "strict" 
from " material" implication and for other basic conceptions of the 
system of "strict implication." A "material" implication represents 
an "If then ** proposition the truth of which is not indepen- 
dent of the truth of its antecedent clause. 


tion of the presentation and our knowledge of the 

Furthermore, when we distinguish one experi- 
ence as illusory, another as presentation of the 
real, what we intend can be nothing other than 
that those sequences of possible experience which 
are implicitly predicted in the concept of the ob- 
ject would fail to eventuate in the one case but 
would be realized in the other. The difference be- 
tween illusion and veridical perception is not in 
the given experience (even though its failure to be 
found there is due to our own inattentiveness or 
other neglect to behave suitably to our interests) , 
else there is not really any illusion. Hence the 
real-ity of any object is known, not by its being 
presented simply, but by judgment or interpreta- 
tion which is predictive. 

Knowledge of objects, then, knowledge of the 
real, involves always two elements, the element of 
given and ineffable presentation, and the element 
of conceptual interpretation which represents the 
mind^s response. We might say that the concep^ 
tual is the formal element, of order or relation, 
and the given is the material or content element. 
But there are here two misunderstandings to be 
guarded against. 

First, the given is not formless in the sense of 
being indefinite. One kind of defimteness which the 
given has its qualitative specificity is too ob- 
vious to need pointing out. Further, it is not form- 


less in the sense that this qualitative and ineffable 
character of it is indifferent for knowledge. If 
there were no correlation in the individual mind 
between the concept and particular qualia, then 
no experience could be the signal of any particular 
meaning. It is also to the point that the implicitly 
predicted relationships, comprised in the concep- 
tual interpretation of what is presented, must be 
such that further possible experience could verify 
or fail to verify them. Without the correlation of 
concept and qualia, no experience could verify or 
fail to verify anything. My presently given ex- 
perience leads me to say that if I should move ten 
feet to the left I should reach the wall. If the 
visual presentation interpreted as "wall" were not 
identifiable by its sensory qualities, or if stepping 
and contact did not have this identifiable qualita- 
tive specificity, then my statement could have no 
meaning. For another person, the sensory qualia 
which would be in point when he saw the wall from 
this chair might be different from mine. But for 
each of us, within his individual experience, if we 
did not correlate certain concepts with certain 
identifiable feelings, there could be no knowledge 
of objects at all. 

The intelligibility of experience consists pre- 
cisely in this ; that between the specific quality of 
what is given and the pattern of its context in 
possible experience there is some degree of stable 
correlation. So that the quality is due to the rela- 


tions and the relational pattern is due to the qual- 
ity. Such stable correlation is not a universally 
discovered fact, all-pervasive in experience. (We 
shall return to this topic in a later chapter.*) Ex- 
perience is completely intelligible only in the sense 
that every experience will exhibit some discover- 
able correlation between presented quality and re- 
lational context. So far as such correlation is a 
fact, it is simply the miracle that an intelligible 
world exists. 

In saying that knowledge consists of the con- 
ceptual element and only points or refers to the 
given, there is no intent to deny that the eventual 
significance of knowledge may be in the quality of 
immediate experience. It is quite possible and 
plausible that the direction of our action and the 
ultimate significance of our attempts to know are 
determined by the value-aspect of experience, 
which is a dimension of, or derivative from, felt 
quality. Knowledge may be, in general, a means 
to some more valuable end which is not knowledge. 
Indeed, so far as the nature of the goods of life 
admits of any ultimate separation of means and 
end, I should suppose this view to be correct- 
Knowledge is pragmatic, utilitarian, and its value, 
like that of the activity it immediately subserves, 
is extrinsic. It has value as an end in itself only 
so far as, in life, the activity is the goal, or at 
least the two cannot be separated. 

*See pp, 348-358. 


Nor do I mean to assert that consciousness is 
essentially relational, though I suspect this to be 
the case. I assert only that it is relation which 
constitutes that intelligibility which is essential to 
knowledge. Whether consciousness of pure im- 
mediacy without conceptual interpretation would 
be below the level of awareness altogether, I hesi- 
tate to judge. But if complete absorption in the 
immediate is not equivalent to unconsciousness, at 
least it is the bourne from which no traveller can 
bring back any intelligible report. 

As was suggested in Chapter II, the mystic and 
the protagonist of "pure perception" will prob- 
ably differ with the account here given primarily 
by their insistence that such absorption in the in- 
expressible immediate is the valuable end. They 
tend to reserve the word "knowledge" for this re- 
lation of the mind to the given which they thus 
esteem it desirable to take. I reserve the word for 
that which is articulate and verifiable, and has a 
significant opposite, "error." So far, the difference 
will be purely verbal. They would agree that ab- 
sorption in the immediate transcends conceptual 
thinking altogether. The further point, that con- 
ceptual thinking, articulation, and the interests of 
action go together, they would also admit. To the 
mystic, negation of conceptual distinctions and 
absorption in the immediate represent the desir- 
able attitude because this solves of problems of 
action, for his world-weariness, by negating them. 


To say, "The identification of mind with the in- 
effable object is the valid end," is the same as to 
say, "The interests of action are nil" 

The motive from which Bergson, who does not 
thus extol the mystic experience, restricts theo- 
retically true knowledge to C pure perception" in 
which practical interests are transcended, is a mat- 
ter which his writings do not make wholly clear. 
Indeed I am minded to ask: What difference 
would it make if his "scientific constructions" and 
"interpretations in the interest of action 9 ' were re- 
named "knowledge" and the object of them "real- 
ity"; if his "intuition" and "pure perception" 
were labelled "esthetic experience" and their ob- 
ject "subjective immediacy"? Apart from some 
moral or religious interest in setting intuition and 
its object higher than science and the scientific 
(and social) object, what point in this ascription 
of "knowledge" and "reality" to pure intuition 
and its object? For Bergson, the scientific (the 
social and the common-sense) interpretation is, 
theoretically considered, a misinterpretation made 
in the interests of practice. But in what sense can 
it be misinterpretation if, follow it however far, 
one never reaches any undesirable denouement 
which could be avoided by refraining? Since con- 
ceptual interpretation serves the interests of ac- 
tion, why this invidious denial to it of the term 
"knowledge" unless action is essentially undesir- 
able and its interests are a low sort? 


He also thinks that conceptual interpretation 
reveals its inadequacy in the end by disclosing its 
internal inconsistency as well as its untruth to the 
given. Tliis is the significance of his resuscitation 
of Zeno's paradoses and of much else in "Time 
and Free Will. 5 ' If he should mean by this that 
prevailing scientific conceptions of time, space, 
motion, etc. are defective and need to be re- 
placed, many will agree with him. But if he means 
as seems to be the case that no scientific con- 
struction in the interests of action and social co- 
operation can escape eventual inconsistency, then 
it would appear to be implied that the interests 
of action are themselves self-frustrating. If the 
significance of conceptual interpretation lies in 
action, then the significance of inconsistent inter- 
pretation must be in self-defeating action. But if 
Bergson should commit himself to the thesis that 
practical action is essentially self-defeating, he 
will indeed agree with the mystic. So far as his 
objection is that scientific conception turns the 
flight of pure duration into objects which are 
static, it may be suggested that the mistake lies in 
supposing that the object denoted by common- 
sense and scientific concepts is thus static, instead 
of something temporally extended and identified 
as a certain predictable flux of experience. 

It is a frequent criticism of the type of theory 
here outlined that it cannot account for our 
knowledge of the past. Knowledge, it is said, is 


here identified with verification, and verification 
comes about by some proceeding from the present 
into the future* Thus the past, so far as it can be 
known, is transformed into something present and 
future, and we are presented with the alternatives, 
equally impossible, that the past cannot be known 
or that it really is not past. 

Without a metaphysical analysis of temporal 
categories it would be difficult to answer this criti- 
cism completely. But I should like to present cer- 
tain considerations which may be put forward 

The first of these is of a general sort which has 
much wider application* The philosophic analysis 
of any object or content of knowledge is not com- 
pletely achieved by even an ideal epistemology. 
The theory of knowledge, to be successful, must 
disclose, for every major type of object, the ratio 
cognoscendL But the achievement of an account 
which should accomplish this would still not abol- 
ish nor make superfluous analyses of other sorts, 
directed to some different problem. In some one 
of the innumerable meanings of the word "is" it 
must be true that a thing is what it is "known as," 
identifiable with its ratio cognoscendi; but it is 
also the effect of its causes, the cause of its effects, 
the organized whole of its physical or other con- 
stituents, and a hundred other significant things 
besides. There has been some tendency in philoso- 
phy since Kant, and perhaps particularly amongst 


idealists, to attach to epistemological analyses a 
kind of exclusive truth, but to do this is to com- 
mit what Professor Perry has called "the fallacy 
of initial predication." If, then, we assert that, 
from the point of view of knowledge, the past is 
so and so, this is not to deny to the past various 
types of significance not included in such an epis- 
temological account* 

It may also be legitimate to observe that the 
criticism in question would have greater weight if 
in general those who urge it were prepared to tell 
us how the past, which is really dead and gone, 
can be known. Epistemological analyses of the 
past, consistent with theories maintained, are con- 
spicuous mainly by their absence. Those who do 
not speak in the interest of such an alternative 
account should not too hastily reject a theory 
which at least begins with an obvious fact (the 
verifiability of the past) because it affects them 
with a feeling of paradox. Paradox is indeed a 
danger-signal, but the trouble it signalizes may be 
in the theory which appears paradoxical or it may 
be in the relatively inchoate character of common 

In general, the past is verifiable. We are prob- 
ably safe in assuming that any satisfactory meta- 
physics will hold that there could not be any item 
of the past which is intrinsically unverifiable. 
Knowledge of the past, like knowledge of any- 
thing else, may be verified only in the present and 


future. But if one suppose this to mean that a 
past event, verifiable as being so and so by certain 
present and future possible experiences, is thereby 
transformed into something present or future, the 
error he commits is in a failure to understand 
what is involved, from the point of view of knowl- 
edge, in assigning temporal locus to an object 
known. We may remind ourselves of the example, 
borrowed from Kant, of the permanent house and 
the impermanent position of the boat in the 
stream, both **known as 5 * certain types of sequence 
in experience, fundamentally alike in their merely 
temporal aspects. That a thing endures through 
a given period is more or less completely verifi- 
able. But if permanence of the thing were justi- 
fiably predicated only of what is permanently ex- 
perienced, there would be nothing permanent. 

The assumption that the past is intrinsically 
verifiable means that at any date after the hap- 
pening of an event, there is always something, 
which at least is conceivably possible of experi- 
ence, by means of which it can be known. Let us 
call these items its "effects.* 5 The totality of such 
effects quite obviously constitute all of the object 
that is knowable. To separate the effects from the 
object is, thus, to transform it into some incog^ 
nizable ding an sich. We may then say, from a 
certain point of view, that the event is spread 
throughout all after time,* much as modern phys- 

*We may neglect the question whether all events are intrinsically 
predictable and, hence, extend through aH preceding time as well. 


ics may say that the field of an electronic charge 
is spread through all space and that this field is 
the electron. That the field of a charge is the 
charge and is throughout all space, does not abol- 
ish the difference between a charge at one point 
and a different charge at another; these are dif- 
ferent ways of being throughout space, a different 
totality of effects. Just so, the conception that an 
event is spread throughout the time "after its 
occurence" will not abolish the difference between 
different events occurring at different times ; these 
will be identifiable through a different totality of 
effects. We must avoid the fallacy of simple loca- 
tion* with respect to temporal as well as spatial 
attributes. In so far as an event, or the existence 
of a thing at a certain date, is intrinsically veri- 
fiable, and is thus spread through all after time 
in its effects, what cognition of it apprehends as 
its presentation and what historical knowledge 
proceeds to verify, is a part of its nature. Or it is 
the "appearance" of the event at the time of this 
verification, much as a presented surface may 
be the appearance of a solid object. Events are 
knowable "after they occur" because their ap- 
pearances or effects are "there" at the later date 
to be experienced. 

For a satisfactory account, it would be essen- 
tial to reveal, by analysis of experience, those pe- 
culiar characteristics by which the pastness of a 

*I borrow this phrase, of course, from Professor WMteheacL 


thing is presently identified ; as would also a simi- 
lar account of the categories "permanent" and 
"future." But for present purposes it will be suf- 
ficient to remark that obviously some kind of 
identifiable marks in presented experience must 
mean the pastness of the thing presented, since 
otherwise the past event could not be distinguished 
from the present. Doubtless one item would be a 
certain kind of unalterability and unresponsive- 
ness to desire and purpose in which respect what 
is present or future would not be thus unalterable. 
This character, like unalterability in general, 
would be verified by proceeding from the present 
into the future in certain ways. In any case the 
past can be known if not completely, at least so 
far as it can come in question for a theory of 
knowledge and it can be known to be past. 
Whatever it is by means of which past fact is 
verified, it is something which is capable of pres- 
ent and future experience. The past is known 
through a correct interpretation of something 
given, including certain given characters which 
are the marks of pastness. If this be paradox, 
then so much the worse for common-sense. 





The history of philosophy since Descartes has 
been largely shaped by acceptance of the alterna- 
tives; either (1) knowledge is not relative to the 
mind, or (2) the content of knowledge is not the 
real, or (3) the real is dependent on mind. 

Kant, and phenomenalism in general, recog- 
nizes the relativity of knowledge, the dependence 
of the phenomenal object on the mind, and hence 
the impossibility of knowing the real as it is in 
itself. Idealism, taking the relativity of knowl- 
edge as its main premise, argues to the unquali- 
fied dependence of reality upon mind by holding 
the alternative that there is no valid knowledge 
of the real to be logically impossible. Realists in 
general seek to reconcile the possibility of know- 
ing reality with its independence of the mind by 
one or another attempt to escape the relativity of 

However, the alternatives accepted are false 
alternatives. This whole historical development, 
so far as it turns upon them, is a mistake. There 
is no contradiction between the relativity of knowl- 



edge and the independence o its object. If the 
real object can be known at all, it can be known 
only in its relation to a mind; and if the mind 
were different the nature of the object as known 
might well be different. Nevertheless the descrip- 
tion of the object as known is true description of 
an independent reality. 

This sounds at first like unintelligible paradox. 
But, as I shall hope to convince the reader, rela- 
tivity of this sort, which is entirely compatible 
with independence, is a commonplace, capable of 
illustration for all sorts of relations which have 
nothing directly to do with knowledge. 

If this position can be successfully maintained, 
then the fundamental premises of phenomenalism 
and idealism fall to the ground, some of the main 
difficulties posed by skepticism are met, and the 
general attitude of common-sense realism can be 
reinstated without attempting to do the impossible 
and avoid the relativity of knowledge. 

The question whether we know reality truly, is 
often made to turn upon the relation between 
presentation, or appearance, and the real object. 
The assumption upon which Descartes set out 
the assumption of the copy-theory is that knowl- 
edge of the external world requires that the sense- 
quale we apprehend should be identically present 
in the object perceived and in the mind when we 
perceive it. Perhaps, remembering the subtleties of 
the critical realists, we ought to phrase this more 


sharply : The validity of sense-perception does not 
depend upon the numerical identity of, e. g. 9 blue 
in the perceiving mind and blue in the object; 
state of mind and property of the object may be 
separate existences. But it is required that they 
be identical in quality ; that the quality perceived 
as blue be in the object just as it is perceived. 
Failing this, the real object is a ding an sich, a 
"something, I know not what." 

That the sense-qualities as perceived are rela- 
tive to the perceiver, and hence subjective only, 
is an old thought. It goes back to Empedocles and 
the Sophists, in the earliest theory of sense-per- 
ception as something produced by the mingled 
motions of the object and of the sense-organ. It 
is the main root of ancient skepticism. The rea- 
sons for It are pretty obvious and historically 
have been well exploited. In the first place, there 
are individual differences of perception, and dif- 
ferences between our own perceptions under dif- 
ferent circumstances. Second, there is the correla- 
tion between differences of behavior power to 
discriminate and compare with differences of the 
receptor organs; from which we conclude a de- 
pendence of sensory qualities upon the senses and 
the nervous system. And third, there is the dis- 
covery of physical phenomena, such as certain 
wave-motions, closely analogous to the stimuli of 
perception but not affecting human senses at all. 
We have, then, good reasons to believe in the 


relativity o the quality of what is given, and the 
limits o what can be given, to the nature and ca- 
pacity of the perceiving subject. This being so, 
can reality be given to us as it is in itself? Is this 
relativity compatible with the truth of immediate 
experience to its object? 

Whether the character of the given is untrue 
to reality in any other sense, it should first be 
admitted that certainly it can be, at times, prac- 
tically misleading. It can be the occasion of er- 
rors of judgment. And the nature and limits of 
the mind or sense-organs may be what gives rise 
to such errors. If I eannot discriminate what you 
discriminate, I may mistake one thing for an- 
other in ways which you do not. In general, for 
every discoverable peculiarity of the subject which 
is reflected in the character of the presented as 
such, there is some imaginable or actual mislead- 
ing which can result from it. The qualia of the 
given are the clue to the applicability or inap- 
plicability of concepts, and set the limits of con- 
ceptual interpretation. Wrong understanding may 
be due to stupidity ; that fact does not here con- 
cern us. But also it may be due to the nature of 
the given. All illusory experience is thus mislead- 
ing. If my poor vision does not enable me to detect 
that one wall of the room is a mirror, or if I 
merely fail to notice this, then I may walk into 
the wall. Even if I do not thus put it to the test, 
my perception leads me to predicate a possibility 


of experience which does not in fact obtain. The 
given, because of its character as given, means for 
me an object which does not in fact exist. With 
greater acuity of vision, I should have avoided 
this error. The reason for it lies in the character 
of presented experience, which is conditioned by 
the limitations of my senses. 

Entirely similar considerations apply in most 
cases of illusion and in all cases of errors of per- 
ception due to limitation or peculiarities of sense. 
But we have so far told only half the story. Such 
errors, we have said, are due to the character of 
the given in the particular case. They are also 
due to the conceptual interpretation which has 
been put upon what is given. In fact they are due 
directly to the conceptual interpretation and only 
indirectly to the given experience* 

The difference between veridical perception on 
the one hand, and illusion and error on the other, 
is not in the nature of what is given except so far 
as this leads to the likelihood of an interpretation 
which is invalid. Such interpretation has to do 
with the relation between the presentation and 
that further possible experience which our classi- 
fying or recognizing or understanding of what 
is presented implicitly predicts. When this pre- 
dicted relation actually obtains, the presented ob- 
ject is recognized for what it truly is. When it 
does not, our understanding of the object is er- 
roneous. Sometimes such mistake amounts to illu- 


sion ; sometimes not. But this is a comment on our 
use of the term "illusion" rather than upon the 
relation between presentation and the real. It de- 
pends upon the degree and importance of our 
failure to apprehend this relation to further ex- 
perience, which in turn depends on the breadth 
of our knowledge and various other circumstances 
affecting our judgment* 

Whether a mirror-image is illusory or veridical 
depends on our previous experience of mirrors, 
or upon the degree of attention which determines 
whether we notice the frame, and so forth. A mir- 
ror-image recognized as such, is veridical percep- 
tion* When I take it for a thing which cas be 
grasped by moving in the direction of the mir- 
ror, it is illusion. If I recognize it as a mirror- 
image but take it as the image of another person 
when it is my own reflection, then we have a per- 
fectly definite and commonplace experience which 
is partly correct and partly erroneous what is 
usually called a "mistake. 95 In these three cases, 
what is given might be identical. Similarly if I 
judge an object ten feet off to be a mile away, it 
is illusion; but if I judge it to be fifteen feet off, 
it is a "slight error." AH sorts and degrees of mis- 
take in the implicit prediction of further experi- 
ence are possible, and readily illustrated ; the flat 
dichotomy into illusion and veridical perception 
ill accords with the facts. In part, this classifica- 
tion reflects the importance of certain interests of 


action which the perception serves. When such in- 
terests are completely thwarted, we have illusion; 
when they are only partly or momentarily 
thwarted, the perception is usually classed as in- 
accurate, more or less mistaken. And in part, our 
use of such terms reflects the fact that the implicit 
prediction of further possible experience would 
often, if explicitly stated, be quite detailed and 
circumstantial, and that such prediction may be 
verifiable in part and in part not. The question 
of veracity is the question how much and how im- 
portant a part of the prediction is valid. There is 
no distinct type of cases, classifiable as "illusion/ 5 
which are the seeing or feeling what is "not true 
to reality 35 as contrasted with normal perception 
which is "seeing things as they are." 

Furthermore, is illusion ever intrinsic to the 
given? Are there cases in which what is presented 
is inevitably mistaken? Obviously the answer is 
that there is no presentation which it is totally 
impossible to mistake and none with reference to 
which it is impossible to avoid mistake. Whether 
one is deceived depends always on the previous ex- 
perience, the breadth of understanding, and the 
wit of the perceiver. The classic "illusions," such 
as the psychologist uses for illustration, are such 
as are pretty sure to deceive us the first time we 
experience them unless we are on our guard. 
Sometimes the so-called "illusory perception" per- 
sists after experience has dispelled our "belief" 


in it. But it is then no more truly illusory than the 
clear and normal visual images which I receive 
since I put on spectacles, but never had in all the 
years before. Whoever has had such experience 
knows that a new correlation of image and dis- 
tance has to be learned. Such an example well 
illustrates the fact thaiTno appearance is intrin- 
sically illusory, because the correlation with ap- 
propriate action and further experience is always 
something which has to be learned. When the 
presentation arouses anticipation or leads to ac- 
tion which is instinctive or habitual, but neverthe- 
less in this instance proves to be ill-judged, the 
perception is classified as illusory or erroneous. 
But it is no more intrinsically so than that of the 
near-sighted man with his first spectacles, who 
misjudges the height of curbings and the distances 
of door-knobs. 

Not all presentations, of course, are presenta- 
tions of real physical objects. There is, however, 
something a little arbitrary in this statement, since 
it is fairly clear that there is never any presenta- 
tion which is not, in one way or another, condi- 
tioned by the existence and nature of certain 
physical things ; and there is thus never any pres- 
entation which is not, for intelligent understand- 
ing, a clue to some physical reality. But there are 
appearances so closely like normal presentations 
of physical things, and so different in their correct 
interpretation, that they may seem to be a special 


class. However, there is no important difference, 
for the theory of knowledge, between such cases 
and those previously mentioned. Only children and 
savages are deceived by dreams, and few are in- 
capable of intelligent discrimination of the merely 

The error which is responsible for this persistent 
notion that some presentations are intrinsically 
illusory, is the prejudice of the copy-theory that 
knowledge of objects means qualitative coincidence 
of the idea and the real. Such a conception is 
meaningless. Knowledge does not copy anything 
presented; it proceeds from something given to- 
ward something else. When it finds that something 
else, the perception is verified. When it fails, or to 
the extent that it fails, we have error or illusion. 
My visual image of the doorknob is my percep- 
tion of its distance, among other things. With my 
spectacles on or off, two somewhat different images 
are signal for the same successful grasping mo- 
tion. If I should try on and become habituated to 
various other kinds of lenses, I should find a great 
variety of such images, any one of which would, 
with experience, signalize the distance to the door- 
knob. That man makes glasses and nature makes 
eyes, does not mean that images seen without ar- 
tificial aid are peculiarly different in their cogni- 
tive significance. If one could try on, as one tries 
on spectacles, all the different kinds of eyes that 
nature has produced, one would find a similar 


variety of visual images, every one of which would 
be, for the owner of those eyes who was accus- 
tomed to them, veridical perception. If one could 
similarly try on all the different kinds of muscula- 
ture and length of arm that nature makes, one 
would acquire a whole museum of sensory "dis- 
tances of the door-knob." Any one of them would 
be perception of reality 3 and none of them would 
copy anything. 

However, this may not touch the question sup- 
posed to be at issue. Are there not presentations 
such that within them, merely as given, there is no 
clue to a veridical anticipation? As presented, there 
is no clue that this is X and not Y. To class it 
as X Is to be wrong ; as Y to be right ; but even 
the most astute and experienced perceiver could 
not find within this momentary experience any re- 
liable clue that it is X and not Y. To admit this 
unqualifiedly might be to exaggerate; we are all 
of us capable of being Sherlock Holmeses about 
familiar imagery. But within limits there certainly 
are such experiences; in fact, all presentations 
have this indecisiveness to a degree. That, how- 
ever, has nothing to do with any supposed intrinsic 
illusion. At the present moment, nothing in my 
perception will enable me to decide whether yonder 
tree is fifty or seventy feet away, whether that 
bird I hear singing is an oriole or a grosbeak, 
whether that weather vane I see is brass or copper, 
whether the breeze that has just sprung up is de- 


cidedly cooler or only a little so. There is nothing 
illusory in these indecisive aspects of my present 
experience, because I have learned not to jump 
to conclusions in such matters. Both veridical per- 
ception and error are judgmental. An infant who 
should see the sights and hear the sounds which 
I do could not possibly be deceived. He does not 
know enough to be wrong. Where there is no in- 
terpretation or anticipation, there can be no error. 
Between such commonplace experiences and those 
which are illusory, there is no intrinsic difference. 
Present experience is a certain clue to very little. 
It enables me to anticipate other experience only 
within limits (the tree is between forty and eighty 
feet away) or with a certain degree of probability 
(the grosbeak's song, as I remember it, is less in- 
terrupted ; this is probably an oriole) . 

All presentation is valid perception when it is 
correctly understood. Understanding is not a mat- 
ter of the qualitative character of the given but 
of the anticipatory attitudes which it arouses. 
What these are, in any particular case, depends 
partly on the characteristics of the presentation 
but equally upon the perceiver, his past experi- 
ence, and his judgment. Since there is no experi- 
ence which is intrinsically incapable of being cor- 
rectly understood and interpreted, there can be 
no presentation which is intrinsically illusory. So 
to speak, any reality does and must appear in the 
way it "ought" to appear to the kind of subject 


who perceives it under the conditions in which he 
perceives it. Between those given experiences which 
are most flagrantly likely to mislead and veridical 
perception, there are all degrees and kinds of in- 
termediaries. No experience can be guaranteed to 
be either veridical or illusory unless the mind of 
the perceiver can be guaranteed. 

In other words, knowledge as valid interpreta- 
tion is independent of the question whether pres- 
entation and real object coincide in quality (if 
that means anything) , because the validity of un- 
derstanding does not concern the relation between 
experience and what is usually meant by "the in- 
dependent object"; it concerns the relation be- 
tween this experience and other experiences which 
we seek to anticipate with this as a clue. In Berke- 
ley's language, this experience is "sign of" other 
experience ; it may be such quite without reference 
to its "copying reality" or even if there be no 
"independent reality" to copy. As the history of 
phenomenalism serves to illustrate, conceptual 
knowledge may be valid provided only there is 
order in experience if experience is lawful quite 
without reference to any further question. 

This disposes of the first point ; that the given 
may be untrue to reality in that it may give rise 
to illusion and mistake. Such error consists in mis- 
understanding or practical misleading ; this is al- 
ways a matter of conceptual interpretation, and 
always it is avoidable. However, this does not dis- 


pose of the question of the relation between the 
phenomenal and the real. Descartes took this same 
position that the content of given experience is 
always veridical, and error is always due to in- 
ference which outruns the percept. But for Des- 
cartes this meant that, when interpretation is 
shorn away, the percept matches the independent 
reality, concerning which God in His goodness 
would not deceive us. 

We may note in passing the total impossibility 
of this representationalist position. We cannot 
both of us see reality as it is when we do not see 
it alike ; and differences of discrimination indubi- 
tably prove divergence of the given in different 
minds. The combination of representationalism 
with realism of the Cartesian type inevitably falls 
before the attack of the skeptic. Indeed, this com- 
bination of doctrines could only survive in an age 
when it was possible to believe that there was some 
normal content of veridical perception, exactly 
shared by the great majority of persons the 
others to be somehow dealt with in a foot-note. 

The relativity of presentation to the perceiver 
can hardly be denied. That this does not affect the 
validity of knowledge, can be established, since all 
knowledge is conceptual or interpretive. But this 
only renders more acute the problem posed by 
phenomenalism: How can a knowledge which is 
relative to the knower's mind and senses be true to 
a reality which is independent? How can the ob- 


ject of knowledge be identified with the real? This 
question resolves itself into two parts; (1) the 
bearing of a knowledge which is relative upon an 
object which has an independent nature, and (2) 
the valid significance of that "independence" 
which may be ascribed to reality. It is the first of 
these which is particularly in point in the dicus- 
sion of phenomenalism; the second has mainly to 
do with certain arguments of idealism. 

We turn to the thesis with which the chapter 
opened* Reality, so far as it can be given in ex- 
perience or known, is relative to the knower. It 
can be apprehended only as it does or would ap- 
pear to some perceiver in some actual or possible 
experience. But that the only character which can 
be attributed to anything real is a character de- 
scribed in relative terms relative to some experi- 
ence does not deny to it an independent nature, 
and does not deny that this nature can be known. 
On the contrary, true knowledge is absolute be- 
cause it conveys an absolute truth, though it can 
convey such truth only in relative terms. 

There is much here that is in no wise peculiar 
to knowledge but has to do with the logic of rela- 
tivity in general. It is equally true of weight, for 
example, that it can only be described in relative 
terms, but that the property of the object, so de- 
scribed, is independent of the particular standard 
in terms of which the description is given. The 
situation in which truth can be told only in rela- 


tive terms, is obviously a common one, if not uni- 
versal. But this relational truth may nevertheless 
be absolute* To put the matter in general terms : 
If relative to R 9 A is X> and relative to S, A is 
F, neither X nor F is an absolute predicate of A. 
But "A is X relative to B" and "A is F rela- 
tive to S" are absolute truths. Moreover they may 
be truths about the independent nature of A. 
Generally speaking, if A had no independent char- 
acter, it would not be X relative to R or F rela- 
tive to S. These relative (or relational) charac- 
ters, X and F, are partial but absolutely valid 
revelations of the nature of A. If we should add, 
"There is no truth about A which can be told 
without reference to its relation to R or S, or some 
other such," we should then have a very good 
paradigm to the relativity of knowledge. 

To make this clear, let us turn to a few simple 
examples of relativity, some of which have nothing 
specially to do with knowledge. 

The size of Caesar's toga is relative to the yard- 
stick.* But if we say, "The number of square 
yards in the toga is determined by the yardstick," 
the statement is over-simple. Given the toga, its 
size in yards is determined by the yardstick ; given 
the yardstick, the number of yards in the toga is 
determined by the toga itself. If the toga had not 
a determinate sizableness independent of the yard- 

*I choose this example partly because no yardstick ever was or 
will be laid on Csesar's toga. I hope the parallel* mil be drawn in 
terms of human minds and past geologic ages. 


stick, or if the yardstick had no size independent 
of the toga, then there would be no such fact as 
the number of yards in the toga; the relation 
would be utterly indeterminate. This independent 
character of the toga, or of the yardstick, is what 
we should be likely to call its "absolute" size. This 
can only be described in terms of some measure, 
though the description will vary according to 
what this measure is. The size of the toga in yards 
is relative to the yardstick, but it is nevertheless 
an independent property of the toga, a true re- 
port of which is given by its correct measurement 
in yards. Thus what is relative is also independent ; 
if it had no "absolute" character, it would have 
no character in relative terms. 

This example leads naturally to another an- 
other sense in which size is relative. One might be 
moved to observe that this conceptual relativity of 
size is something which goes round in a circle. The 
toga is of so many yards ; a yard, so many feet. 
But how big is a yard or a foot? Eventually this 
goes back to something like the king's foot, which 
is a fact of the same order as the toga. A size is 
relative to other sizes; but some size must be an 
absolute so-bigness, immediately apprehended, or 
there is no size at all. This would be to maintain 
the eventual reference of the concept to something 
immediate. But it is well to note in passing that, 
except in precisely such relative terms, the abso- 
lute so-bigness of the king's foot is also an absolute 


Size as an absolute and immediately given so- 
bigness Is quite similarly relative when this size 
is attributed to the object. As Berkeley put it: 
How big is a mite's foot? As big as it looks to the 
mite or as big as it looks to us? We here confront 
a relativity of sense-experience which concerns its 
supposed truth to the real object. Size, as per- 
ceived, varies with distance from the perceiver. 
And there is no possibility of perceiving size at 
all except at some distance. Perceived size is a 
function of two terms, distance and X. The dis- 
tance being fixed, differences of perceived size are 
attributable to differences in X. The perceived 
size is, so to speak, the value of the function. This 
is a function of two variables ; its value, perceived 
size, is not determined by distance alone; it de- 
pends also on X. Distance being specified, and the 
value of the function, perceived size, being given, 
X is thereby determined. Distance being known, 
the perceived size is a true revelation of X, which 
we may call the independent size of the object. 

If it be asked, "But what precisely is this in- 
dependent size in any intelligible terms," we can 
carry our mathematical analogy one step further. 
(It is, in fact, a little more than an analogy, for 
the logic of functions is not confined to mathe- 
matics,) The independent size, X 3 of the per- 
ceived object is the integration of the function, its 
perceived size or perceived sizes, over the whole 
range of the other variable, distance. That is, all 


the perceived sizes, at different distances, belong to 
or are parts of the objectively real size of the thing 
perceived. The analogy holds good, further, in 
that from the value of the function, perceived size, 
for any given distance, its value for other dis- 
tances are predictable. If, now, we remember that 
the conceptual interpretation of the immediately 
presented as the size of an objectively real thing, 
is precisely such implicit prediction, from its per- 
ceived size at this distance, of perceived sizes at 
other distances (among other things), we shall 
observe that the "independent size of the object" 
is precisely the content of a correct concept by 
which its size as presented is understood. For such 
a correct conceptual interpretation, any one of its 
perceived sizes is a true revelation of this inde- 
pendent property. 

If any one ask for an absolute size which per- 
ception or knowledge could copy or be true to in 
any fundamentally different sense, I can only say 
that the meaning of his inquiry escapes me, and 
I believe it escapes him also. 

It is obvious that what is here pointed out for 
size, holds for properties in general. Just as size 
may be in terms of the king's foot or the platinum 
bar in the Bureau of Standards, so color, for ex- 
ample, may be determined by reference to the 
sun's spectrum or the color-pyramid. This is con- 
ceptual relativity. Thus, in turn, we may seem to 
be thrown back on color as perceived, the visual 


quality just as we perceive it. But this, as a prop- 
erty of the real thing, is something which can 
hardly be supposed to represent simple coinci- 
dence of mind and object, because color, as per- 
ceived, varies with illumination. Except in light 
of some candle-power color cannot be seen at all, 
but the perceptual content itself varies with va- 
riation of the candle-power. Does this mean that 
we never see color as it is? Or that we see it as 
it is only at some standard illumination, arbi- 
trarily determined? Or that we always see it as it 
is, when we see it at all, if what we see enables us 
to predict our altered visual experience of the ob- 
ject under other conditions? 

Similarly for shape. Conceptually a shape is 
relative to other shapes. It can be described only 
in relation to standard shapes, such as square and 
round or by analysis into elements of shape, such 
as angles (measured by reference to a standard) 
and linear measure (obviously again relative). 
And shape as immediately presented configura- 
tion, if referred to the object, is relative to per- 

The logic is the same throughout. Relativity is 
not incompatible with, but requires, an independent 
character in what is thus relative. And second, 
though what is thus relative cannot be known 
apart from such relation, still the other term or 
terms of the relation being given, all such relative 
knowledge is true knowledge of that independent 


character which, together with the other term or 
terms of this relationship, determines this content 
of our relative knowledge. The concept, or con- 
ceptual interpretation, transcends this relativity 
precisely because what the concept comprises is 
this relational pattern in which the independent 
nature of what is apprehended is exhibited in ex- 

This being so, the nature of the fallacy com- 
mitted by phenomenalism becomes apparent. From 
the relativity of knowledge to the mind, it argues 
to the impossibility of knowing the independent 
real. This is as if the question about the size of 
Caesar's toga were to be answered : "Its size in our 
yards is so and so ; in terms of some other measure 
which other creatures might apply, it would be 
different. Apart from yards or some other mea- 
sure, size has no meaning. So you see that the 
real toga in itself is something outside the cate- 
gory of size. Whether it can have size at all or, 
if so, what that size would be> we can never know. w 
The premise is correct. The conclusion non 

It may seem that our illustrations are not pre- 
cisely to the point since neither distance nor de- 
gree of illumination nor angle of perspective, etc., 
is a property of the perceiving mind or of the 
sense-organs. But such illustrations have the ad- 
vantage of making it clear that the logic of rela- 
tivity is unaltered whether the object in question 


is an independent real, supposedly beyond or be- 
hind its appearances altogether, or is recognized 
to be the merely phenomenal object. The penny 
whose apparent size and configuration are rela- 
tive to distance and perspective is recognized to 
be phenomenal and wholly knowable whether it is 
admitted to be independently real or not. If the 
relativity of its various appearances to perspec- 
tive, distance and so on, does not defeat the possi- 
bility of our knowing the phenomenal object, how 
can relativity to mind, the logic of which is point 
for point identical, defeat the possibility of know- 
ing the independently real object? 

The one ground on which it might be urged 
with some show of reason that the relativity of the 
content of knowledge to the mind prevents true 
knowledge of the real, is that the nature of one 
term in this relation the mind itself is not 
known. That is, it might be said that we cannot 
stand outside ourselves and critically bound our 
own limitations. The elliptical appearance of the 
penny it may be urged conveys true knowledge 
because I know my angle of vision, and know how 
this appearance would vary as my perspective 
was altered. If all objects were seen from one 
angle only as all objects are perpetually viewed 
from within the limitations of the human mind 
then the relativity of perceived shape to perspec- 
tive would lead to a confusion of configuration as 
(always) perceived with an absolute shape, which 
it is not* 


There is much here that is worthy of careful 
attention, though the point is not sufficient to 
establish the phenomenalisms conclusion. 

In the first place, it needs to be remarked that 
the criticism, as put, overlooks the fact that I can 
not know my own angles of vision except through 
those same given configurations, and the altera- 
tions of them, by which I know the shapes of 
things. I learn to understand both objective shape 
in general and the phenomenon of perspective in 
general when I learn how to introduce order into 
the given phenomena of perceived shape by treat- 
ing them as functions of two variables, different 
perspectives and different objective shapes. And 
the analogue holds: I can know my own mind 
through its commerce with objects and only so 
just as I know objects through their commerce 
with mind. I learn to understand both objective 
reality in general and the general character of my 
own human mind when I learn how to introduce 
order into the procession of given presentation by 
treating this experience as a function of two vari- 
ables, the subject and the object. To revert to the 
mathematical terms, the data of appearance are 
the values of the function, cognition. This is a 
function of two variables, mind and object. I 
know the object by an integration of its appear- 
ance over the range of the other variable, mind 
or "the subjective conditions*" (For example, as 
has been pointed out, objective change is divided 


from permanence of the thing by integrating with 
respect to those changes, in experience as given, 
which are due to "my own activity.") And I know 
mind by an integration of this function over its 
whole range (or the widest possible range) of 
variation in the objective. That the phenomenal- 
ist treats mind as transcendent is a fallacy which 
is correlative to his treatment of the independent 
object as beyond knowledge. 

We must, further, distinguish between the no- 
tion that unrecognized limitations of the human 
mind would mean any deceitfulness or erroneous- 
ness of knowledge a failure to accord with the 
true nature of the real and the quite different 
notion that such limitations would mean a corre- 
sponding degree of ignorance of reality. When 
this distinction is drawn, the whole point of phe- 
nomenalism, as regards the relation of mind to 
independent reality, will be found to be lost; be- 
cause when we grant to his arguments the utmost 
which can be granted, the conclusion to which they 
point is that our knowledge of the independent 
object is veridical but partial, not that it is un- 
true to absolute reality* 

Our analogy may be of further assistance here. 
It is true that if we were restricted to one angle 
of perspective, one distance, etc., this would lead 
to limitation of knowledge. If we were restricted 
to perception at five feet, whether we knew it or 
not, that would mean a real limitation of our 


knowledge, because we could not understand or 
predict the systematic variation of perceived size 
with distance, which is an additional insight into 
the nature of that independent JST, the size of the 
object, which is one term of the relation which 
determines size as perceived. Because we are able 
to see things at a wide range of distances, we learn 
to predict from our image at any one distance 
the appearance of the thing at other distances. 
Though the momentary perception is limited to 
a single distance, the breadth of previous experi- 
ence, and knowledge of this momentary condition, 
enable us to transcend the momentary limitation. 
A permanent limitation could not be thus tran- 

However, we must not confuse limitation of 
knowledge with misrepresentation or mistake. Any 
sort of limitation of sense-organs or mind which 
should be reflected in perception, so far from 
meaning that we do not perceive things as they 
are, means that in certain respects we are freed 
from all possibility of error and are fatally cer- 
tain to perceive things as they are, though to per- 
ceive and understand only part of what we other- 
wise might. The penny which looks elliptical may 
deceive us into thinking that it is elliptical, pre- 
cisely because we are capable of viewing things 
from other angles than the present one. If we 
were limited to just one angle of vision, we should 
be restricted in our knowledge but we should 


thereby be freed from all possible mistakes of 
perspective. Similarly, if the image I have of a 
mite's foot should register on the retina of an 
intelligent mite, it would lead to error. But the 
limitation which prevents him from seeing his 
foot as it looks to me, at the same time prevents 
him from suffering certain illusions about his feet 
which would otherwise be possible. 

That we humans do not have senses which reg- 
ister directly the whole known range of harmonic 
motions, means that there is much of reality which, 
until we learned to call upon various indirect 
modes of observation, was beyond our knowledge. 
And our inability to imagine how certain ranges 
of vibration might register upon sense-organs 
which should be sensitive to them, is as much a 
limitation as the blind man's inability to imagine 
color. But just as blindness does not condemn a 
man to false perception or even false interpreta- 
tion (although it does make it practically neces- 
sary to run more risks and hazard judgment in 
the absence of desirable clues) , so in general sub- 
jective limitations cannot render knowledge un- 
true to its object. At most they only mean greater 
ignorance and consequently greater likelihood of 
false judgment. The exigencies of life to be met 
remain just as numerous; the basis of judgment 
is more meager ; hence error will probably be more 
frequent. Yet it remains true that no experience, 
however limited, is or can be intrinsically misrep- 


How much of reality we can grasp, doubtless 
depends upon our human limitations. The relativ- 
ity of knowledge to the mind means such limita- 
tions, but it does not mean that the real object is 
a ding an sich. Unless we grossly suppose that 
what humans can know is all there is to know, such 
limitation can lead to no untruth of our knowl- 
edge to reality. In this matter, as in many others, 
theory seems to suffer from a tendency to ex- 
tremes to hold either that the reality we know is 
all there is, or else that we cannot truly know any. 
The golden mean seems both modest and sensi- 
ble. Knowledge has two opposites, ignorance and 
error* The relativity of perception may mean 
ignorance or it may not. If I can observe things 
from every angle, the restriction to one perspec- 
tive at a time will not mean necessary ignorance, 
especially since other perspectives can be pre- 
dicted from the present one. But if perception 
were restricted to a single angle, that relativity 
would mean ignorance. This will be true for limi- 
tations of the mind in general. 

Ignorance of whatever sort increases the likeli- 
hood of error, because it means that in practice 
we must go forward on grounds of judgment less 
sufficient. But the given itself is never misrepre- 
sentative ; always it is true revelation of the real, 
however partial. The notion that it can be untrue 
to the real reflects both a misapprehension of the 
significance of the truth of judgment and the old 


and meaningless fallacy that the function of sen- 
sory awareness in knowledge is to provide a quali- 
tative replica of the independent object. Igno- 
rance, however great, cannot make of reality a 
ding an slch; it does not vitiate such knowledge as 
we have, and that knowledge is of the independent 

Idealists are wont to draw an opposite sort of 
conclusion from the same general considerations ; 
that is, to urge that the conception of reality be- 
yond all human power to know is meaningless, 
precisely because of the relativity of knowledge in 
general to the mind. It is true and important, that 
we can conceive no kind of reality whatever except 
in terms of some possible experience; but the 
idealist fails to do justice to our human power to 
transcend, by indirect methods, limitations of di- 
rect experience. He should be careful not to deny, 
by implication, that the blind man can believe in 
color. In a sense, the blind man does not know 
what he believes in ; nevertheless he meaningfully 
believes in something that he can neither perceive 
nor imagine. It is very likely true that if we con- 
jecture that reality has aspects forever beyond 
the reach of human beings, we must do so by the 
metaphor of some mind differently organized than 
our own. But when we know that other humans 
have greater auditory range than ourselves, and 
can reasonably suppose that insects possess senses 
which directly register stimuli which we do not, 


what prevents us from conceiving that there are 
ranges of the real beyond the direct apprehension 
of any human, or even of any animal that hap- 
pens to exist? It is further true that we must have 
some sort of conception in terms of which to as- 
scribe reality of any type; otherwise the ascrip- 
tion of reality itself means nothing. But if the 
idealist puts his challenge in the form, "How can 
we know there is a kind of reality we cannot 
Jcnow?" the different significance of the word 
"know** in its two occurrences needs to be consid- 
ered. Mr. Russell has pointed out that we know 
there are numbers which nobody will ever count. 
To "know" a number is to know whether it is odd 
or even, prime or factorable, etc., or at least to 
be in position to determine this. To know that 
there are numbers not thus known, is to know a 
principle of the relation of every number to others, 
by which further counting is always possible, and 
hence to know that some numbers will always be 
uncounted. Similarly we may know (or have good 
reason to conjecture) that there are certain sys- 
tematic relations in reality by which what is di- 
rectly perceptible to us is connected with what is 
not. If it is a question how we are to conceive 
what should be beyond experience, then we may 
warn the idealist that he ought to be careful of 
this point lest he spoil his own argument. He has 
a final metaphysical interest in the power of hu- 
man beings to transcend their own finitude. This 


is a case in point; in fact, it is the general case. 
We transcend our own limitations in terms of 
what we should or might experience if . That is 
the nature of possibility in general. We transcend 
actual experience in terms of possible experience. 
But we transcend the actually given only as we 
abstract from something which is a fact; the na- 
ture and extent of such abstraction determines 
the degree and kind of possibility which is in 
question. There is no limit to the number and kind 
of restrictions of human experience which we can 
thus speculatively transcend. 

Can I see both sides of this coin in my hand? 
At this moment when I am looking at the obverse, 
I cannot see the reverse side. But if I should turn 
it over, I should see the reverse. Seeing both sides 
of the coin is both possible and impossible pos- 
sible on condition, impossible without that condi- 
tion. Taking my limitations severely enough, any 
possibility can be ruled out except the actuality. 
The meaning of a possibility which transcends the 
actual lies in the truth of some "If then" propo- 
sition, the hypothesis of which is contrary to fact. 
As we progressively transcend the limits of the 
actual by our a if" the possibility in question be- 
comes a more and more attenuated sort, but at no 
point, while our "If then" proposition still has 
any meaning, can we say this possibility is not 

Now, as has been pointed out, the ascription 


of any reality beyond immediate experience re- 
quires and represents such affirmation of the pos- 
sible. To repudiate all such transcendence is to 
confine reality to the given, to land in solipsism, 
and in a solipsism which annihilates both past and 
future, and removes the distinction between real 
and unreal, by removing all distinction of veridi- 
cal and illusory. The ascription of reality is, then, 
the affirmation of possibility, and the kind of real- 
ity ascribed conforms to the nature of the possi- 
bility affirmed. As we progressively transcend our 
actual limitations, the reality conceived becomes 
more and more abstract and undetermined in its 
nature. But at no point, while our hypothetical 
statement still retains a vestige of meaning, is the 
conception of the corresponding real completely 
empty. The conception of other minds, different 
from our own, is a perfectly meaningful "if" by 
which we go one step beyond the actual situation 
of the blind man with respect to color. The reality 
thus speculatively affirmed is one degree more 
blank and dubious, but the existence of such real- 
ity still has meaning. 

It is not, however, with the idealist's argument 
upon this point that we are principally concerned. 
The more important consideration for us is that 
meaning of "independence" of the object which 
is compatible with the relativity of knowledge. 
The idealist argues from this relativity of knowl- 
edge to the mind to the conclusion that the object 


is completely mind-dependent. In this, he misin- 
terprets the nature of relativity and forgets the 
possibility that the object as known may be coin- 
cidently determined by two conditions and thus 
relative to both. 

Since the idealistic argument is different from 
the phenomenalisms, let us vary our illustration 
of relativity in general. 

An alpha-particle is shot out from a radium- 
atom, describes a certain path, and is arrested by 
a screen. The mass of the particle, its velocity, and 
its time of flight, are all such as to be inexpres- 
sible except with reference to some observer or 
frame of motion. The determination of each of 
these properties will vary for different relative 
motions of the observer and the system containing 
the alpha-particle. But does this variation mean 
that velocity and mass are not properties of an 
independent reality which can be observed in these 
various ways? If that were true, then there would 
be no objective physical difference between an 
alpha-particle and a beta-particle or a rifle bul- 
let. The physical identification of the object would 
depend altogether upon the relative motion of it 
and the observer. But if the properties were not 
in some sense determinate independently of any 
frame of motion, then they would not be determi- 
nate in relation to the observer. Specify the rela- 
tive motion of two systems, and these properties 
must have fixed values, representing the physical 


nature of the object. 3?or any observer, in relative 
motion to it such and such, an alpha-particle shot 
out by an atom of radium will have a determined 
velocity x and mass y. 

That is, there are certain properties of the ob- 
ject, as an independent reality, which can only be 
described in terms of some observer or frame of 
motion. But specify this relationship and the true 
description is thereby fixed. What is it that de- 
termines this? It cannot be the relative motion 
already specified. It is fixed by the objective real 
character of the thing. If human observers, di- 
rectly or indirectly, see the motion of the particle 
and measure its mass and velocity, then what they 
observe will depend on them (their relative mo- 
tion). But this condition in terms of them being 
specified, what they observe will depend on what 
they observe. This "what 55 is a determinate thing 
in some sense independent of the relative motion, 
though describable only in terms of some such re- 
lation. Under all conditions, this independent 
"what," along with the relative motion, enters 
into the determination of what is observed. 

In other words, the observed mass, velocity, etc., 
of the objectively real thing is a function of two 
variables, the relative motion of observer and ob- 
served and the independent character of the thing 
observed. Specify both of these and the value of 
the function, observed mass or observed velocity, 
is completely determined. Specify either, and the 


value of this function then depends upon the 
other. Specify neither, and the function is then 
completely indeterminate. 

Say to the physicist : "An alpha-particle is shot 
out from a radium-atom. What is its velocity and 
mass?" He will reply: "Your question is not 
strictly answerable until you specify the relative 
motion of the system and the observer. But since 
it is an alpha-particle, I can tell you the mass and 
velocity for any observer you please. The fact 
that it is an alpha-particle determines a series of 
velocities and masses relative to observers in all 
possible motions relative to the path of it." But 
say to him: "The relative motion of two systems, 
A and 5, is one hundred thousand miles a second, 
directly toward each other. Something moves in 
system A. What is its mass and velocity as mea- 
sured by an observer on J5?" Obviously he will re- 
ply by asking what you are talking about. Or he 
might answer : "Until you specify what the thing 
and its state are for some other relative motion of 
observer and observed (say, rest), I cannot tell 
you its mass and velocity relative to the motion 
of the two systems which you mention." 

Now the objective reality, alpha-particle, is 
identified by certain observable properties, mass 
and velocity amongst them. Suppose it were not 
identifiable by molecular combinations into which 
it enters or by any other properties which are not 
affected by motion, so that it could be known only 


through characteristics which are thus relative. 
We may then imagine ourselves to make rejoinder 
to the physicist: "What I mean is indescribable 
except in terms which are relative to the motion 
of the observer. How, then, can I tell you the na- 
ture and state of the thing I mean in any terms 
that are not dependent on the motion of the ob- 
server?" He might reply: "There is a systematic 
connection between masses, velocities, etc., in 
terms of one relative motion and in terms of any 
other. So it is unnecessary for you to try to an- 
swer in other than relative terms. The description 
in terms of amy relative motion, if that motion be 
specified, will be a sufficient description of the na- 
ture and state of the thing. But surely you do not 
expect me to deduce the nature of the thing from 
the single condition of its motion relative to the 

The parallel in the case of the relativity of the 
object known to mind is obvious. "Thing as 
known" is a function of two variables ; it depends 
on the mind, but also it depends on the thing. 
This thing can only be described in terms of its 
relation to some (actual or hypothetical) mind. 
But this does not alter the fact that if "thing as 
known" were not determined by a condition which 
is independent of the mind, it would not be deter- 
mined at all. The parallel holds also in that the 
thing as known in one relation to a mind (say, 
from one perspective) enables us to predict its 


character as known in other such relations. So far 
as this is the case, its (relative) nature as known 
is a sufficient determination of its nature in gen- 
eral, or independent of any particular relation of 
this sort. 

The fallacy of idealism lies in arguing: "The 
nature of the thing as known always depends on 
the nature of the mind. Therefore the object can- 
not exist or have character independent of the 
mind. 55 This is as if one should argue "The mass 
and velocity of an alpha-particle always depends 
on its motion relative to the observer. Therefore 
it can have no mass and velocity, and cannot ex- 
ist, independent of this relative motion." In one 
sense of the word "independent" it is true that the 
mass and velocity of an alpha-particle has no 
meaning independent of its relative motion. And 
in a strictly parallel sense, it is true that the na- 
ture of the object, independent of the knowing 
mind, is undetermined; and independent of any 
and every mind, is meaningless. But there is no 
need for us to trip over the ambiguities of the 
word. The mass and velocity of an alpha-particle 
at least has two independent conditions; its mo- 
tion relative to the observer is only one of them. 
We cannot argue from "dependent on its relative 
motion" to "completely determined by its relative 
motion." Similarly we cannot argue from the fact 
that it is meaningless to try to describe a thing 
out of relation to mind to the quite different 


thesis that the real object known is completely de- 
termined by the mind which knows it. If it should 
be said that while the object is not determined by 
its particular relation to a particular mind, it is 
determined by relation to mind in general, we may 
revert to the analogy once more. Mass and veloc- 
ity apart from relation to some frame of motion 
is always undetermined; but it is not determined 
by such relation in general (if that means any- 

To revert to a previous illustration, the idealis- 
tic argument may be parodied: "The size of Cae- 
sar's toga is relative to the yardsick or to some 
other standard of measure. No size without a yard- 
stick. The size of things is through and through 
yardstickian. To be sure, the fallible yardstick in 
my hand may not determine size in general, but 
the yardstick in the Bureau of Standards deter- 
mines both my yardstick and all sizes that there 
are. It creates size." 

If the mind were the only condition of the thing 
as known, then the nature of the mind being speci- 
fied, objects in general would be completely de- 
termined. One could say, "Given human mind 
possessed of such and such organs and interpret- 
ing data in such and such categories, what will be 
the reality it knows?" And there would be an an- 
swer in general and in particular. 

Idealism has often boggled over the fact that 
it could not deduce the particular content of ex- 


perience and knowledge. The questions, "Why do 
I have just this experience? Why do I find just 
this reality and no other?" must have an answer* 
Either that or it must be recognized that the par- 
ticularity of experience is itself an ultimate if 
inexplicable datum; that the given is a condi- 
tion of reality independent of the mind. Berke- 
ley, of course, has his reply to this question: There 
is a reality, God, independent of my mind, which 
is responsible. The post-Kantian idealists, not 
sharing Berkeley's empiricism, have either neg- 
lected this problem or, like Pichte, have said that 
it is no part of the business of philosophy to de- 
duce the particular. But he fails to face the ques- 
tion: Granted the idealistic thesis, can the par- 
ticular be deduced?* Philosophy, he might rightly 
claim, is not interested in the fact that I now see 
a blue blotter or that there are elephants in Af- 
rica. But his claim is that all the conditions of 
experience and reality are contained in mind. Out- 
side of minds is nothing which could determine, or 
help determine, what minds know. If that be true, 
then mind being specified, not only the form or 
general character of knowledge but also the con- 
tent in all its particularity, must be determined* 
It would still not be the business of philosophy to 

*ScheIling, however, acknowledges the justice of the challenge and 
seeks to meet it with amazing results. Starting from the Fichtean 
premise, A = A, he deduces eventually the electrical and magnetic 
properties of matter! System, d. transcendentalen Idealismus, sammi. 
Werke (1858), Bd. I, 3, pp. 444-450. 


make the deduction for each particular item, be- 
cause the particular items would not be of gen- 
eral interest. But it is a matter of general inter- 
est that all such items are thus deducible if that 
be a fact. We might except some inductive proof 
of this deducibility the deduction of the ele- 
phants in Africa as an illustration- If the mathe- 
matician should tell us, "All the facts of physics 
can be deduced from the system of quaternions," 
but should reply to our request for a deduction 
of the law of gravitation by saying, "Particular 
physical facts are of no interest to the mathema- 
tician and no part of his business, 55 we should 
draw our own conclusions. 

That idealism may argue that reality is ex- 
clusively mental or spiritual by maintaining that 
the condition of this particularity is another 
spiritual being, we are not here concerned. Such 
argument (or dogmatic assertion) is metaphysi- 
cal and is, or should be, quite distinct from the 
argument from the relativity of knowledge. 

It is a much more important consideration, I 
believe, that unless the content of knowledge is 
recognized to have a condition independent of the 
mind, the peculiar significance of knowledge is 
likely to be lost. For the purpose of knowledge is 
to be true to something which is beyond it. Its in* 
tent is to be governed and dictated to in certain 
respects. It is a real act with a real purpose be- 
cause it seeks something which it knows it may 


miss. If knowledge had no condition independent 
of the knowing act, would this be so? 

It is most important to discover precisely what, 
in terms of knowledge, can be meant by "the in- 
dependence of the object." Whether the subject- 
object relation is universal in reality or not, 
clearly the answer must be compatible with its 
universality in "knowledge. It must, further, be 
independent of any supposed qualitative identity 
of the content or perception with the object, since 
such identity is probably meaningless and in any 
case is unverifiable. 

It may be asked, "What would it mean for a 
mind to know an object, when the supposition of 
the qualitative identity of given content of per- 
ception with the object is ruled out?" The an- 
swer, in terms of the theory here presented, will 
be clear: It means that we are able to interpret 
validly certain given items of experience as sign 
of other possible experience, the total content of 
such further possible experience, related to the 
given in certain categorial ways, being attributed 
to the object, as constituting what we know of it 
and what we mean by attributing reality to it. If 
this conception seems to leave us in the air about 
the "nature of the object," let us first inquire 
what further question it is to which we seek the 
answer. We might find that there is no such fur- 
ther question which is meaningful. 

In terms of experience and knowledge, the in- 


dependence of reality its independence of the 
knowing mind means, first, the givenness of 
what is given ; our realization that we do not cre- 
ate this content of experience and cannot, by the 
activity of thinking, alter it. Second, it means the 
truth of those "If then" propositions in which 
the process of possible experience, starting from 
the given, could be expressed. The "if" here de- 
pends upon our own active nature for its mean- 
ing, as has been pointed out, but the content of 
the "then" clause, and the truth of the proposi- 
tion as a whole, are things with respect to which 
the knowing mind is not dictator but dictated to. 
I may confront the given with different attitudes 
and purposes ; I may be differently active toward 
it and, starting from it, I may proceed into the 
future in different ways. But what I should then 
find; what eventuations of experience are gen- 
uinely possible ; that is something independent of 
any purpose or attitude of mine. These, I seek 
correctly to anticipate in my present interpreta- 
tion of the given. If they do not obtain in reality, 
my present "knowledge" is false* Whether they 
obtain or not, is determined independently of my 
mind. If not, then it is not determined at all, and 
knowledge and error are, both of them, purely 
subjective and meaningless. 

Third, the independence of reality means the 
transcendence by reality of our present knowl- 
edge of it; it means that I can ask significant 


questions about ray object which have an answer 
when that answer is something which I cannot 
give. In terms of experience this means that, 
starting from the given in certain ways, I can 
safely predict the accrual of something the par- 
ticular nature of which I cannot now determine. 
For example, if I examine the contents of this 
drawer, either I shall find a piece of chalk or I 
shall find none. So much I know; but I do not 
know now and cannot discover merely by tak- 
ing thought which of these alternatives I should 
find true. There is that in the object which I do 
not now know ; I know something to be determined 
in reality which is neither implicitly nor explicitly 
determined in my knowledge of it. This, and all 
similar questions I could ask and could not now 
answer, witness the independence of my object. 

If the idealist should find that there is nothing 
in such "independence" which is incompatible 
with his thesis, then it may be that between a suffi- 
ciently critical idealism and a sufficiently critical 
realism, there are no issues save false issues which 
arise from the insidious fallacies of the copy-the- 
ory of knowledge. 



The position so far arrived at emphasizes the 
fact that there is no knowledge of external reality 
without the anticipation of future experience* 
Even that knowledge implied by naming or the 
apprehension of anything presented* is implicitly 
predictive, because what the concept denotes has 
always some temporal spread and must be iden- 
tified by some orderly sequence in experience. 
Hence we are inevitably confronted as any the- 
ory of knowledge must be with the problem of 
Hume's skepticism: Are there any necessary con- 
nections in experience? Can conceptual order, 
which is of the mind, be imposed upon a content 
of experience which is independent and not yet 
given? This is the problem of the a priori. 

There is no knowledge without interpretation. 
If interpretation, which represents an activity of 
the mind, is always subject to the check of fur- 
ther experience, how is knowledge possible at all? 
That the interpretation reflects the character of 
past experience, will not save its validity. For 
what experience establishes, it may destroy; its 
evidence is never complete. An argument from 



past to future at best is probable only, and even 
this probability must rest upon principles which 
are themselves more than probable. For the valid- 
ity of knowledge, it is requisite that experience in 
general shall be in some sense orderly that the 
order implicit in conception may be imposed upon 
it. And for the validity of particular predica- 
tions, it is necessary that a particular order may 
be ascribed to experience in advance. 

Thus if there is to be any knowledge at all, 
some knowledge must be a priori; there must be 
some propositions the truth of which is necessary 
and is independent of the particular character of 
future experience. But traditional conceptions of 
the a priori have broken down, largely because 
the significance of its necessity and its indepen- 
dence have been misconstrued. 

"Necessary" is an ambiguous word ; its contra- 
dictory is, in one meaning, "contingent," in an- 
other "voluntary." The necessary character of 
a priori truth, which is genuinely opposed to its 
contingency, has been confused with some psy- 
chological or other necessity, which the mind is 
tinder, of accepting it. What contradicts neces- 
sary truth must be genuinely impossible to hap- 
pen. But it is not therefore impossible to believe. 
What is a priori does not compel the mind's ac- 
ceptance. It is given experience, the brute-fact 
element in knowledge, which the mind must accept 
willy-nilly. The a priori represents the activity of 


mind itself ; it represents an attitude in some sense 
freely taken. That we elicit some formula as a 
principle means that we take it as forbidding 
something or denying something which in some 
sense has significance. That which is utterly inca- 
pable of any alternative is utterly devoid of 
meaning. The necessity of the a priori is its char- 
acter as legislative act. It represents a constraint 
imposed by the mind, not a constraint imposed 
upon mind by something else. 

And the a priori is independent of experience, 
not because it prescribes a form which experience 
must fit or anticipates some preestablished har- 
mony of the given with the categories of the mind, 
but precisely because it prescribes nothing to the 
content of experience. That only can be a priori 
which is true no matter what. What is anticipated 
is not the given but our attitude toward it; it 
formulates an uncompelled initiative of mind, our 
categorial ways of acting. Truth which is a priori 
anticipates the character of the red; otherwise, it 
would possess no significance whatever. The real, 
however, is not the given as such, but the given 
categorially interpreted. In determining its own 
interpretations and only so the mind legis- 
lates for reality, no matter what future experi- 
ence may bring. 

If we are to understand this nature of the a 
priori, traditional misconceptions must first be 
cleared away. In general these are three: (1) that 


the a priori is distinguished by some psychologi- 
cal criterion such as the "natural light" or some 
peculiar mental origin such as innateness; (2) 
that it is distinguished by some peculiar mode of 
proof , or logical relation to experience in general, 
usually called "presupposition" ; (3) that the 
a priori legislation of mind can not apply to ex- 
perience unless what is given in experience is 
already limited or determined in some consonant 
fashion; that the validity a priori of our cate- 
gorial interpretation requires also a priori modes 
of our receptivity or intuition. 

The first of these need not detain us long: in- 
nate ideas are a dead issue. Psychological tmde- 
niability, even if it exist, would not be proof of 
truth. It is entirely conceivable that the animal 
man should be so organized that certain fallacies 
should be peculiarly impelling to his mind. And 
historically it is observable that what has ap- 
peared undeniable and been accepted as axiomatic 
over long periods of time may still be false. Nor 
is it implausible that there should be truths which 
are a priori, having a warrant not drawn from 
the particular character of particular experi- 
ences, which nevertheless should be grasped only 
with difficulty and not specially impressive to most 
men. Moreover, if the criterion of the a priori 
were a certain impulsion of the mind, then there 
would be no difference amongst truths on this 
point. As Bosanquet has pointed out, all discov- 


ered truth lays upon the mind the same impulsion 
to belief; this character belongs to all proposi- 
tions once they are established* 

The source of this rationalist conviction that 
the a priori must have some peculiar psychologi- 
cal warrant, is fairly easy to make out. Univer- 
sal propositions drawn from experience are con- 
tingent and problematic unless they have some 
prior warrant* Knowledge which is certain can 
not be grounded in the particulars of experience 
if it is to apply to particular experiences in ad- 
vance; it can only come from the possession of 
some universal by which the particular is implied. 
Nor can these universals be reached by generaliza- 
tion* Hence there must be universal truths which 
are known otherwise than through experience. 
Such universal propositions cannot be logically 
derived unless from other such universals as prem- 
ises. Hence there must be some universal truths 
which are first premises logically underived and 
representing an original knowledge from which we 
start. Such propositions must be axiomatic, self- 

However, this notion of innate truth or self- 
illuminating propositions is not particularly con- 
sonant with rationalistic theory* The essence of 
human reason is a mode of thought, not a particu- 
lar content; it has to do with the validity of con- 
clusions, not with original premises natively pos- 
sessed* Post-Kantian rationalism realizes this and, 


influenced no doubt by Kant's deduction of the 
categories, turns from psychological compulsion 
as the ground of the a priori to a conception of 
logical necessity. First, or highest, principles are 
no longer regarded as immediately evident, but 
are now supposed to be distinguished by a pecu- 
liar criterion of proof. They are "necessary pre- 
suppositions" of some class of more particular 
facts, of science, or of experience in general. 

The meaning of "presupposition" here is far 
from clear; probably it has no single meaning, 
and no discussion could be altogether just to the 
variety of its uses. But in general, what seems to 
be intended is the designation of certain princi- 
ples as logically prior to that which "presupposes" 
them, with the added thought that what is thus 
prior is thereby proved to have the character of 
necessary truth necessary, that is, if facts of 
science or experience in general are taken for 

So far as this is what is meant, the fallacy com- 
mitted by the notion that principles can be proved 
true a priori by being presupposed by science or 
experience, is so simple that it is extraordinary 
that it could ever have gained currency. Correctly 
speaking, what is logically prior to a fact or 
proposition will imply that fact or proposition, 
but it will not, in general, be implied by it. In the 
language of mathematics, if A is logically prior to 
S 9 then A must be a sufficient condition of B or 


at least one of a sufficient set of conditions; but 
"sufficient condition" must not be confused with 
"necessary condition." Physics presupposes math- 
ematics in the sense that it exhibits particular in- 
stances of general mathematical principles, while 
mathematics contains no necessary reference to 
physics. In the same sense, all the special sciences 
presuppose logic. But if what is presupposed in 
this sense be regarded as thereby established or 
proved necessary, the fallacy involved is easily 
detected. If I assert that two feet and two feet 
are four feet, I do not thereby commit myself 
to the proposition 2 + SS = 4. It is required only 
that this be true of linear measure. Gases under 
pressure or living organisms might for all that 
is here in question be governed by very different 
laws. The particular fact does not even require 
that there should be any general laws of mathe- 

There can be little doubt that this fallacy has 
played its part in traditional conceptions of the 
a priori* Presuppositions, so called, are always 
general in their import. That which presupposes 
them is more particular. Now A is not a necessary 
condition of B unless "A is false" implies "B is 
false," which is the same as to say that B implies 
A. Hence no general principle is a necessary con- 
dition of any particular fact or proposition un- 
less that particular implies the general principle. 
And even if this should be the case, it would be the 


particular and not the general which was, so far, 
logically prior and the original premise. 

If we avoid this fallacy and take "A presup- 
poses .B" to mean "A is necessary condition of 
B," i. ., "S implies A" then we should be so 
cluttered up with presuppositions that the fine 
glamor of the word would be lost. Presupposi- 
tions would be truly necessary conditions that is, 
relatively necessary ; necessary if but the neces- 
sary conditions of any proposition are as numer- 
ous as the things that it implies. The necessary 
conditions of any particular fact of experience 
are merely its logical consequences. Obviously, it 
is not intended to reduce a priori principle to the 
status of one among the numerous consequences 
of the particular fact. Furthermore, the only ne- 
cessity which could thus be established would be 
relative to the fact in question. If that fact be 
contingent, as the particular content of experi- 
ence is, then its presuppositions will, unless other- 
wise supported, share precisely that contingency. 

The metaphysical respect in which presuppo- 
sitions have been held reflects the vast influence 
exercised by the geometry of Euclid upon historic 
rationalism. This respect is, of course, entirely 
justified; but along with it went a conception of 
geometrical method and of deduction in general 
which, although perhaps inevitable to an earlier 
day, is quite unwarranted. According to this view 
the logically first principles, or presuppositions, 


are self-evident axioms which, through the proc- 
ess of deduction, shed the glory of their certainty 
on all the propositions deduced from them. But 
here it is to be observed that the ground of cer- 
tainty of the first principles has nothing to do 
with their logical priority. The criterion of their 
truth is their self -evidence or undeniability ; they 
loan their indisputable character to their conse- 
quences instead of deriving it from the fact of 
being logical foundation of these consequences. 
If they were not self-evident they could derive no 
certainty or necessity from the fact of being thus 
presupposed. The connotation of the phrase "a 
priori" was fixed in terms of this ancient concep- 
tion according to which all systematic knowledge 
was supposed to find its warrant through deduc- 
tive derivation from such self-evident beginnings. 
Literally connoting "by deduction" it came to 
mean "necessarily true" because only such first 
principles of deduction as were taken to be neces- 
sary or self -evident were then acceptable. 

To-day, however, when this conception of de- 
duction has been given up in mathematics and 
elsewhere, when "postulate' 5 or the colorless 
"primitive proposition" has replaced the self- 
evident axiom, when non-Euclidean geometries 
have been recognized to have precisely the same 
logical structure as Euclid, and when it has been 
shown that various sets of postulates may give rise 
to the same deductive system, we have less than no 


excuse for retaining the notion that a presupposi- 
tion is more certain than its consequences. Where 
the body of facts which a deductive first principle 
implies is considerable and well-established, and 
there are no implications of it which are known to 
be false, the presupposition gains that kind of 
verification which particulars can give to general 
principles that is, the partial and inductive veri- 
fication of it as an original hypothesis. But to 
regard a presupposition as established by what 
presupposes it, except in this inductive sense and 
with the same contingency as the consequences of 
it, has not even the warrant of historical confu- 

The traditional rationalist conception that 
metaphysical first principles can be shown to be 
logically indispensable, or that what is logically 
prior is thereby proved to be certain or self-evi- 
dent, is one to which the actual structure of logical 
and mathematical systems lends no support. In 
genuinely rigorous deductive systems, as these are 
understood today, "logically prior" means only 
"deductively more powerful" or "simpler." The 
supposed necessity, or logical indispensability, of 
presuppositions most frequently turns out to be 
nothing more significant than lack of imagination 
and ingenuity. The plurality of possible begin- 
nings for the same system, and the plurality of 
equally cogent systems which may contain the 
same body of already verified propositions but dif- 


f er in what else they include, dispel the notion of 
indispensability in what is logically prior, 

A less important but equally persistent fallacy 
is the notion that at least some necessary truths 
can be established by the fact that to deny them 
is to reaffirm them that they are implied by their 
own contradictories. It is wise to walk cautiously 
here, because the logical facts are quite complex. 
The most frequently offered illustrations of re- 
affirmation through denial are not even good cases 
of a proposition implied by its own denial. For 
example, the fallacy of arguing from the undeni- 
able existence of thinking to the self which does 
the thinking vitiates Descartes's use of the "I 
think." But quite apart from that, the man who 
should assert "I am not thinking," so far from 
contradicting himself, would give the best possible 
evidence of the truth of his statement. The propo- 
sition, "I am not thinking," does not imply, "I am 
thinking*" It may be that the attitude of will 
which we suppose to underlie the making of any 
assertion is such as to be incompatible with the 
admission, U I am not thinking," so that we may 
be sure that whoever could make such a statement 
would find himself at cross purposes. But the rea- 
son for this is contained neither in the proposition 
nor in any implication of it. There is here no logi- 
cal inconsistency whatever. 

Other examples of the supposedly self -contra- 
dictory the statement of Epimenides the Cretan 


that "All Cretans are liars, 5 * etc* have been so 
f requently discussed in current literature that con- 
sideration of them may be omitted here* Most of 
these commit what Mr. Russell calls a "vicious 
circle" fallacy by ignoring the systematic am- 
biguity of type which characterizes such proposi- 
tions. As a corrolary, the supposed "necessity" 
which attaches to their contradictories is equally 

However, there are propositions which are genu- 
inely implied by their own denial, and hence 
propositions whose denial leads to their reaffirma- 
tion. And all such belong to the class what may 
quite reasonably be called "necessary." (I should 
omit further consideration of such logical tech- 
nicalities, which must be getting boresome to the 
reader, except that this particular point will be 
of some importance later on.) The curious fact is, 
about such genuine examples of propositions im- 
plied by their own denial, that they are not thus 
proved true. To see that this is so, we must first 
examine the nature of reaffirmation through de- 
nial. Whoever asserts a self-contradictory propo- 
sition does not in one and the same breath affirm 
and deny the content of his assertion. He affirms 
it in fact ; he denies it by implication only. Or to 
put it otherwise; he affirms it, and the question 
whether he also denies it is the question of what 
his assertion implies. 

Now the most obvious illustrations of such 


propositions whose denial genuinely implies them 
come from the field of logic ; in fact, they all be- 
long to logic when that subject is interpreted in 
the rational way as including all purely formal 
truth. And the content of logic includes all prin- 
ciples of inference. 

Whoever, then, denies a principle of logic, may 
either draw his own inferences according to the 
principle he denies, or he may consistently avoid 
that principle in deriving his conclusions. If one 
deny a principle of inference, but inadvertently 
reintroduce it in drawing conclusions from his 
statement, he will indeed find that he has contra- 
dicted himself and admitted what originally be 
denied. But if he denies a principle of inference 
and consistently reasons in accordance with his 
own statement, he need incur no self-contradiction 

It is a fact that for one who stands within a 
given system of logic, the denial of one of its prin- 
ciples will imply the principle itself. But this sig- 
nifies nothing more profound than the fact that 
deductions in logic are inevitably circular.* In de- 
ducing our theorems of logic, we must make use 
of the very principles which the deduction is sup- 
posed to demonstrate. If then, I use "bad" logical 
premises but "good" logical reasoning, I shall 

*0mitting from consideration the development of logic, as a purely 
abstract system, by the "operational" instead of the "posfrulatory" 
method. These omitted considerations serve to strengthen, not to 
weaken, what is here set forth. 


contradict myself 3 quite as surely as if I use two 
premises which are mutually inconsistent. Perhaps 
an example here will be of assistance* Take the 
law of contradiction in the form, "That X is A 
and X is not A 9 is false." Its contradictory will 
be, "X is A and X is not A" Let us take this last 
statement as a premise and draw the inference 
from it. 

(1) "X is A and X is not A" implies its latter 
half , "X is not A." 

(2) "X is not A" implies "It is false that X 
is A." 

(3) "It is false that X is A" implies "That X 
is A and X is not A y is false." (Just as " 'Today 
is Monday' is false" implies "That today is Mon- 
day and it is raining, is false.") Thus from the 
denial of the law of contradiction we have deduced 
the law of contradiction itself. But we have done 
so only because, though denying it in the premise, 
we have reintroduced it in step (&) of the rea- 
soning. If we had, consistently with the premise, 
refused to take step (S), we should never have got 
any such conclusion. 

Every good or correct logic, then, will be such 
that its principles are undeniable without contra- 
diction; the denial of any one of them leads to 
formal inconsistency. But this is true only because 
so long as we remain within our system of logic, 
we shall use the very principle in question in draw- 
ing inferences from the denial of it, and thus beg 
the question of its truth. 


A good logic must be circular. But what should 
lead any one to suppose that this character be- 
longs exclusively to systems of good logic? Ap- 
parently those who set store by the "reaffirmation 
through denial" have committed the fallacy of 
illicit conversion; they have reasoned: "A logic 
whose principles are true will give their reaffirma- 
tion through denial. Therefore, whatever princi- 
ples meet this test must be true." 

All logic and pseudo-logic is similarly circular* 
A little ingenuity suffices to construct a bad logic 
in which, reasoning badly according to our bad 
principles, we always get consistently bad results. 
And if we deny one of these principles, still by 
sticking to our bad method of reasoning, we can 
reaffirm the bad principle in conclusion.* Since 

*One family of such systems consistent in their own terms, and 
such that the denial of any principle lends to its reaffirm ation as a 
consequence is determined by the presence in the system of the 


where p. q, etc. are propositions, and p < 0. represents ^ p implies g, 
or "if p is asserted, q may be asserted." This proposition allows of 
two distinct meanings of p <q> neither of which coincides with the 
usual one; and the properties of this relation may be further speci- 
fied in a variety of ways. Some of the systems in this family might 
be regarded as "good" logic, but most of them are "bad." Such a 
"bad" logic may be developed logistically from the following formal 

A. ( p) =*p (Def. of p* the denial of p) 

B. (p<p) 

C. (P<?) <(/7<r<-*9 

D. [p - 

F. (p <q) <( p < t 
Postulate F is obviously false as a general law of implication* It is 
interesting that postulate B seems to exclude the possibility that any 
proposition should lead to its own denial as a consequence, yet if P 
beanyprincipleof thesystem, we can prove that P<( P < P) 
Hence the assertion of ( P) leads to the assertion ( P < P ). 


a bad logic, whose principles are false, may still 
be such that the denial of any one of these prin- 
ciples will lead to its reaffirmation, it follows that 
the test of "reaffirmation through denial" does not 
in logic prove the truth of the principle thus re- 

It should be added, to avoid misunderstanding, 
that in spite of what has just been said, the test 
of self-criticism or circularity is a valuable test of 
any deductive development of logic. That the 
principles proved are precisely the principles used 
in the demonstration of them, is here a matter for 
congratulation. That the method of our proof co- 
incides with the result of it, is a test of both method 
and result. It is not a test of truth, however ; it is 
a test of formal or methodological consistency. 
The error of taking self-criticism to be a test of 
logical truth lies in overlooking the fact that a 
thoroughly false logic may still possess this merely 
methodological consistency. 

One further hit of explanation seems required 
also. I do not mean to say that there are no neces- 
sary propositions. Whoever takes a given logic to 
be true will find its principles undeniable without 
contradiction (L e^ in his logic) and therefore 
necessary. Some logic is true, and hence some log- 
ical principles are necessary. The point is simply 
that the truths of logic are not proved by any 
such procedure since, as proof, it always begs 
the question. 


Precisely the point which. I wish here to make 
is that logical "necessity" has here no connotation 
of the inescapable. What is a priori is not true 
because the mind is so constituted that it finds 
such truth unavoidable; however fantastic or 
practically negligible any alternative supposition 
may be, there still are such alternatives, which 
may be self -consistent. Doubtless what is funda- 
mental, as logic is fundamental, has its roots in 
the nature of the human mind, but not in such 
wise as to be either self-evident or the only self- 
consistent possibility. If it should be such that it 
must be assumed or it cannot be proved, that, so 
far from proving truth, would be a character 
which it shares with delusions and absurdities. 
There will still be alternatives of assumption in 
the presence of which the mind is uncompelled. 
Whatever was genuinely imposed upon the human 
mind would not be a priori; it would have just 
that brute-fact character which distinguishes the 

It is here that rationalist conceptions, by their 
confusion of logical and psychological, fall into 
further difficulty. The a priori is recognized as 
not being given as the content of experience is 
given. But if the a priori have psychological self- 
evidence or inescapability of any sort, then it must 
be absolute datum in some sense or other. Either 
the mind would find these truths belonging to it 
as soon as it became conscious, with sufficient clear- 


ness, of its own possessions, or it would acquire 
them at some particular date in some particular 
moment of illumination, and would recognize that 
the ground of this new realization was not pre- 
viously there in the mind. In the former case, the 
a priori would belong to the mind only in the 
sense that the individual body does; the infant 
would find it as he finds his ears or finds that he 
can move his arms. It would be a commonplace of 
reality, but it would have no higher character than 
that of uniformly evidenced fact. There would be 
no guarantee of it beyond the guarantee of uni- 
form experience up to date. On the other alterna- 
tive, it would have the character which belongs to 
such illumination as may be received from an ex- 
ternal and authoritative source, and the truth of 
it would depend upon some sanction superior to, 
or at least independent of, his own mind. Yet it 
is of the essence of rationalism to recognize the 
a priori as a peculiar possession of the mind itself, 
in a sense not compatible with either of these con- 

This point of the relation of the a priori to the 
mind, is really of prime importance, for upon it 
depends that assurance, superior to the assurance 
we can have of generalizations from experience, 
that nothing future experience can reveal will 
falsify it. Whatever experience may bring, the 
mind will be there ; whatever belongs to the mind 
itself is assured in advance. This is the one point 


upon which all conceptions which recognize an 
a priori have agreed. The conception which re- 
tains this significance and avoids the fallacies and 
contradictions pointed out above, is one which out- 
rages traditional ideas, but at bottom it is simple 
and, I think, can be made obvious. The a priori 
has its origin in an act of mind; it has in some 
sense the character of fiat and is in some respects 
like deliberate choice. The a priori is a peculiar 
possession of mind because it bears the stamp of 
mind's creation. And the criterion of creativity is 
not inevitability but exactly its opposite, the ab- 
sence of impulsion and the presence of at least 
conceivable alternatives. But I dare not press a 
point of view so novel until there has been further 
consideration of points which historical concep- 
tions and problems will serve to exemplify. 

In particular, if the a priori is to be thus con- 
ceived as made by mind, shall we not fall into an- 
other difficulty : How, then, shall we know that it 
can be imposed upon a reality which is indepen- 
dent? It is here, of course, that we find the grounds 
of skepticism in general and of Hume's in par- 
ticular. The human mind, by its nature and by 
the manner of its activity, imposes certain inter- 
pretations upon experience. Every such interpre- 
tation would, if valid, limit the character of re- 
ality and the possibilities of future experience. We 
can have no assurance that such limitations char- 
acterize the independent real or bound what future 
experience may bring* 


It is for tliis same reason that Kant recognizes, 
in addition to the categories, another a priori ele- 
ment in knowledge and makes the distinction of 
phenomenal and real. The content of experience 
is limited by the forms of intuition, which are im- 
posed not by the active interpretation of the mind 
but by the passive modes of its receptivity.* The 
categories are subjective modes of the mind's in- 
terpretation or synthesis of the content of intui- 
tion. How, then, can we be assured that they will 
be valid of experience in general? An indispensable 
part of Kant's answer is that the object in ex- 
perience must itself be subjective or phenomenal. 
It must be limited by the very fact of being ex- 
perienced in such wise as to make universally pos- 
sible the mind's modes of categorial synthesis. 
That which can not validly be thought under the 
categories can not be given in intuition. Thus the 
objects of knowledge are the objects of experience. 
The limitations of thinking are also the limita- 
tions of sensing; the possibility of knowledge is 
assured by the fact that experience is not of the 
independent real but of phenomena already in- 
formed by our receptivity. 

*It is difficult, of course, to interpret Kant, with any assurance of 
accuracy, upon this point. In the Transcendental Aesthetic the distinc- 
tion of forms of intuition as modes of our receptivity from any mode 
of mind's activity, seems sufficiently clear. In later sections, refer- 
ence to the ^'synthesis of apprehension" and such passages as the 
one headed, "Of the a priori grounds of the possibility of experience" 
raise doubts about the principle of this separation. But certainly the 
division between those conditions which the mind actively imposes 
on the object and those which it passively imposes in intuition must 
be there, else the whole procedure of the Critique falls to the ground. 


This manner of meeting the skeptical difficulty 
is both unnecessary and impossible. It is impos- 
sible because if there were conditions imposed upon 
experience by our receptivity, mind could not 
recognize them as its own. At most it could only 
conjecture that they belonged to it and not to the 
nature of the independent real, or to that portion 
of reality to which experience so far had been con- 
fined. Lacking any certain criterion by which the 
limitation of the content of experience could be 
ascribed to mind, such conditions would appear 
simply as limitations of what was given, whose 
continuance in all future experience would be as 
problematic as any empirical generalization. And 
this answer to skepticism is unnecessary, because 
mind may limit reality (in the only sense which 
the validity of the categories requires) without 
thereby limiting experience. The active interpre- 
tation by mind imposes upon given experience no 
limitation whatever. 

Every beginning student of Kant asks sooner 
or later, "But how does Kant know that phenomena 
are not things in themselves? 55 And the only an- 
swer that can be given is that if what could be 
experienced were limited only by what existed to 
be experienced, then the limits of experience could 
be discovered only through experience itself. Any 
conclusion rgarding them would then be probable 
only, since it would be argument from past to 
future. If the limits belong to reality and not to 


the mind, then knowledge of them a priori is not 

Perhaps this answers the question why Kant, 
consistently with the rest of his procedure, must 
distinguish phenomena from things in themselves. 
But it omits the real question how we can know 
that the limitations of experience are due to the 
mind and are not simply those of an independent 
reality which experience reveals. If there are limi- 
tations of experience which are imposed not by 
the activity of thinking addressed to the given, 
but before that given is given, or in its being 
given, how shall we distinguish what mind is re- 
sponsible for from what independent reality is 
responsible for? This can only be done either by 
knowing the unknowable reality or by some cri- 
terion of what mind is responsible for in given 
experience. This must take the form, "Even if it 
existed to be experienced, we could never experi- 
ence X (let us say, non-Euclidean space )." And 
this reminds us of another objection that the be- 
ginning student makes: "How do we know that 
we shall keep on having the kind of mind we have 
and not wake up tomorrow in a noil-Euclidean or 
timeless world?" Probably the answer is that we 
do not know this; that Kant in fact supposes it 
possible (since he believes in immortality) ; but 
that it is useless to mix our problems in this way ; 
for discussion of the validity of mundane knowl- 
edge and of science, the general character of hu- 


man experience is a datum in the sense of setting 
the problem of explanation* But this does not an- 
swer the question how, if we should wake up in so 
novel a world, we should know that the change was 
in us, in the forms of our receptivity, and not 
merely in external reality. 

We cannot conceive any limits of possible ex- 
perience in general. Or to speak more exactly; 
the limits of the possibility of experience are the 
limits of meaningful conception. "Possibility of 
experience" is ambiguous, and "conception" as 
that word is ordinarily used, is ambiguous in a 
parallel fashion. The sense in which I can con- 
ceive a non-Euclidean reality is the sense in which 
I can give meaning to an "If then** proposi- 
tion in which the "if ** states some intelligible con- 
dition and the "then" ascribes some content which 
supposedly would be experienced under the con- 
ditions of that "if." How abstract and fantastic 
such an "if" may become, without losing mean- 
ing altogether, and how tenuous the speculatively 
conceived reality which is thus ascribed, we have 
seen in the previous chapter. But in precisely the 
same sense that "reality" is ascribed, in that sense 
some possibility of experience is predicated. We 
can conceive limits of human experience only by 
conceiving the possibility of an experience which 
we do not have. When the possibility of experi- 
ence is speculative, the reality in question and the 
limitations which it transcends are equally specu- 


lative. Where that possibility has some basis in 
actual experience, a limitation may be known, but 
it is known by generalization from experience and 
the prediction of its continuation in all future 
experience has precisely and only such assurance 
as may attach to empirical generalizations. We 
cannot know a priori and with absolute certainty 
that any limitation of experience will be perma- 
nent. I can and do conceive my individual experi- 
ence as limited by certain personal peculiarities 
such as near-sightedness and range of auditory 
sensibility. But I discover this limitation by com- 
parison with other persons and by conceiving 
something, which I do not have, as possible experi- 
ence intelligibly related to the kind I do have. 
Whether such limitation is permanent or tem- 
porary, is merely a question for empirical science 
to answer; the oculist or aurist may tell me. I 
conceive all human experience to be similarly lim- 
ited by human sensibility, as compared with that 
of some other animal, or perhaps only some 
dreamed-of being whose sensibility should be af- 
fected by stimuli which affect none of ours. We 
cannot literally imagine such experience any 
more than the blind man can imagine red ; never- 
theless it is not beyond our powers of conception. 
It is an identical proposition that no conceivable 
experience or reality is beyond our powers of con- 

Let us not forget the issue which is in question. 


The skeptic asks how we can know that modes of 
conception or understanding or interpretation 
which are of the mind can be validly imposed upon 
an independent reality and all future experience. 
The Kantian answer is that the object of knowl- 
edge is not independent reality but phenomena, 
which are limited by human modes of receptivity, 
and that these, in the nature of the case, will hold 
for all future experience. This answer is unneces- 
sary because we have a much simpler one before 
us ; it is an identical proposition that no conceiv- 
able experience or reality is beyond our powers of 
conception. What is beyond our powers of con- 
ception has no meaning; the word which is sup- 
posed to denote it is a nonsense syllable. Experi- 
ence does not need to be limited in order that we 
should be able to understand it ; we can understand 
anything in one way or another. 

Furthermore, whatever is understood is in some 
sense or other conceived as possible of experience. 
The very manner in which we attach a meaning, 
for example, to "non-Euclidean space" is assur- 
ance that if we should experience it we could un- 
derstand it. The conception of non-Euclidean 
space is a fairly definite one, and the sense in 
which it could be understood, if experienced, is 
definite in the same measure* Other conceptions of 
what transcends actual experience are more tenu- 
ous; the ^^possibility 35 of experience which is in 
question is more abstract and the meaning of the 


hypothesis is vague in like degree. But what is 
absolutely and in every sense beyond the possi- 
bility of experience is likewise beyond all mean- 
ing. An absolute and a priori limitation of ex- 
perience could not be known. The kind of limita- 
tions which I attribute to myself or to human 
beings in general is the kind which I associate with 
sense-organs or some biological characteristic; 
and whatever evidence of them there may be is 
empirical. We know that we shall never experi- 
ence directly certain ranges of vibration because 
we have no sense-organs which are affected by 
them. But we know this a posteriori only, and with 
whatever probability attaches to the continued 
correlation between certain sensory apparatus and 
certain modes of stimulation, or between our sub- 
jective experience and these particular organs. 
That our experience will persist in having this 
limitation, is probable only, however fantastic the 
alternative. If it be asked, "How do we know a 
priori that there are certain limits which will 
characterize all future experience," the answer is 
that we can have no such knowledge. We shall ex- 
perience what we shall experience. We might ex- 
perience anything you please, imaginable or un- 
imaginable, which can be phrased at all. And if 
we experienced it, we should proceed to under- 
stand it, either by finding a consistent categorial 
interpretation or by condemning it as hallucina- 
tion (which, after all, would be a categorial un- 


derstanding of it.) The only limitation which need 
be imposed upon possible experience in order that 
it may be brought under the categories is the limi- 
tation to what can be understood. The alternative 
to what can be understood cannot even be phrased. 
And what is limited only by nonsense syllables is 
not limited at all. 

It may be objected: Exactly the contingency 
which is supposed to be ruled out by the Kantian 
conception is the possibility of a fantastic ex- 
perience in which all that order upon which sci- 
ence relies, all my knowledge and modes of inter- 
pretation, would be worthless. Unless experience 
is limited a priori, what rational ground of as- 
surance have we that knowledge may not be thus 
invalidated? The answer is in part that which the 
queen gave in the episode of the wishing carpet: 
"If this were real, then it would be a miracle. But 
miracles do not happen. Therefore I shall wake 
presently. 95 We have no absolute assurance in ad- 
vance that our experience at some particular time 
in the future will meet the criteria of physical 
reality. But we are sure in advance that if it does 
not, it will not be experience of the physically real. 
Kant creates an artificially difficult problem for 
himself by his use of the term "experience" as if 
experience and the phenomenally real coincide- 
Did the sage of Konigsberg have no dreams 1 In 
fact, this procedure is quite usual. We first for- 
get all that part of experience which is under- 


stood by being classed as dream or illusion, and 
think of "experience 9 ' as something to the content 
of which reality must be assigned. We then pro- 
ceed when confronted with the skeptic's chal- 
lenge to terrify ourselves with the possibility of 
an "unintelligible" experience such as any pro- 
saically-minded person would immediately dispose 
of, if it were reported to him, with the verdict 
"dream" or "insane imagination." The categories 
are required to cover the totality of experience 
only if the categories "dream," "illusion," "hallu- 
cination," are included. An a priori principle of 
interpretation is not required to bring all experi- 
ence within that category whose principle it is. 
Precisely what it expresses is the criteria of real- 
ity, of a certain type such as the physical. Its 
universal applicability to experience is satisfied 
if whatever experience does not conform to the 
criteria in question can be repudiated as not real 
(e. g. 9 not physical reality). Obviously any ex- 
perience is intelligible if the absence of certain 
types of order mark it as unreal and, therefore, 
not in question. And could any one's experience 
be understood without repudiating much of it as 
non- veridical? A priori principles of categorial 
interpretation are required to limit reality; they 
are not required to limit experience. The con- 
tingency of illusion, dream, or even of insanity 
may be real possibilities of future experience ; that 
has nothing to do with the validity of the cate- 


In this connection, it is also of importance to 
raise another and related question: Has it ever 
been claimed, or could it reasonably be claimed, 
that knowledge of the particular can be a priori? 
In the sense that this particular can, and must be, 
subsumed under some universal of which knowl- 
edge is possible a priori; Yes. But in the sense 
that mind, confronted with a given content of 
experience, can with absolute certainty refer it to 
its proper category, and thus interpret what is 
now given in such wise that no further experience 
could invalidate that interpretation ; No. The par- 
ticular phenomenon may always be non-veridical 
or the subject of mistaken apprehension; and 
whether it be such, we look to further experience 
to reveal. The usual phrasing of our dictum with 
respect to the particular might be such as : "This 
material object must have mass," or "The sum of 
two sides of this triangular plot must be greater 
than the third. 55 But the empirical object is al- 
ways such that we are capable of being deceived 
about its "true nature. 55 This presentation may 
not be a real material thing ; the plot may not be 
truly triangular. Strictly such knowledge of the 
particular is always complex: "This is a material 
object and material objects must have mass. 55 
"This plot is triangular, and the sum of two sides 
of a triangle must be greater than the third. 55 The 
first half of these represents the subsumption of 
the given under a category ; the last half, a prin- 


ciple of that category. Clearly it is the principle 
only which is a priori. The interpretation by 
which the given is referred to a category is al- 
ways such that it may possibly be erroneous; its 
validity can never be known with certainty, in- 
dependent of all further experience. That the 
categories condition experience in the sense of im- 
posing on its content an irrevocable order and 
connection, or an interpretation not subject to 
doubt in the particular case, it is not possible to 
believe. The only sense in which categorial in- 
terpretation can be a priori is the sense that the 
principle of this interpretation is not subject to 
recall even if, in the particular case, what is given 
should fail to conform. That is a priori which we 
can maintain in the face of aU experience no mat- 
ter what. In the case of an empirical law, a mere 
generalization from experience, if the particular 
experience does not fit it, so much the worse for 
the "law." But in the case of the categorial prin- 
ciple, if experience does not fit it, then so much 
the worse for the experience. 

The question of the possibility of knowledge a 
priori, is not the problem: How can we know in 
advance that experience which should not con- 
form to our categorial principle is impossible? It 
is the problem : How do we know in advance that 
if it does not conform to our principle it will not 
be veridical, or will not be real in the category 
which is in question? The former question can 


have no answer unless by some impossible dog- 
matism about the limitation of experience by a 
mind which is itself above or behind experience, 
and hence unknowable* And even this hypothesis 
of transcendent mind does not assure the perma- 
nence of its conditions unless by some further dog- 
matism which assumes its continuity unchanged. 
The latter question has an obvious answer: We 
know that any experience which does not conform 
to our categorial principle will not be veridical 
because the principle states the criteria of reality 
of that categorial type. 

If it be asked further: "How do we know, in 
these terms, that we may not be presented in ex- 
perience with what will not fit into any category 
and thus be wholly unintelligible," the answer is 
in part by reference to that systematic ambiguity 
of the term "reality" which was pointed out in 
Chapter I. What is not reality of one sort is 
reality of another ; what we do not understand in 
one way, we shall understand in another. The 
subsumption of the given under the heading 
"dream 9 * or "illusion" is itself a categorial inter- 
pretation by which we understand certain experi- 
ences. Even "the unintelligible" is a sort of cate- 
gory, a temporary pigeon-hole in which items are 
filed subject to later classification when we have 
some further light on them or it becomes more 
imperative to understand them. It would be a 
hardy soul who would insist that no content of 


experience is unintelligible at the moment when it 
is given, or that there is any time-limit on the un- 
intelligibility of particular items of experience* 
The notion that all content must be immediately 
and absolutely intelligible and categorized, in be- 
ing admitted to experience, is just one of those 
respects in which rationalistic theory is too pretty 
to be true. 

It might even be suggested, without going be- 
yond the bounds of plausibility, that the assump- 
tion that nothing can be finally and absolutely 
unintelligible, is a sort of ideal of reason, or rep- 
resents a willingness to bet on our capacity to 
triumph over any apparently chaotic character 
of experience and reduce it to some kind of in- 
telligible order. Our dictum that no experience 
can be intrinsically unintelligible, is saved from 
being falsified by experience in general by the 
fact that it sets no time-limit on our efforts to 
understand, and hence no failure can be final. As 
a report of our actual dealings with the given, the 
generalization "All experience is understood 55 
would be a bit absurd. 

It is, however, more important and more just 
to observe that intelligibility is always a matter 
of degree. Nothing is completely understood. And 
some partial interpretation is always possible. 
That very repudiation by which the non-veridical 
is ruled out from a certain category, is Ttself such 
partial interpretation and represents a beginning 


of our understanding of 'the experience in ques- 
tion. Thus the ascription of intelligibility and 
unintelligibility is always relative relative to our 
present powers and relative to those interests 
which make interpretation in some particular way 
momentarily important or desirable. With refer- 
ence to our present understanding, all experience is 
both. The notion that, in categorizing the given, 
the mind understands it completely and has done 
with it for all time, is an unwarranted assumption 
which the Kantian point of view seems to incul- 
cate without explicitly making. And it is implausi- 
ble as soon as it is mentioned. 

Hence if we take the problem of the a priori to 
be concerned with our foreknowledge of absolute 
limits of the possibilities of all future experience 
as such, the question how we can have such knowl- 
edge has no valid answer. But if we take it to re- 
late to our knowledge in advance of the princi- 
ples to which all veridical experience must con- 
form, it has an obvious one. The principles of 
categorial interpretation are a priori valid of all 
possible experience because such principles ex- 
press the criteria of the veridical and the real. No 
experience could possibly invalidate them, because 
any experience not in conformity, which might be 
evidence against them, is automatically thrown 
out of court as not veridical in that category, and 
hence not pertinent to them. Knowledge of such 
a priori principles requires only reflective self- 


consciousness because it is simply knowledge of 
those criteria which we apply in classifying ex- 
perience in one or another fashion, interpreting it 
in one or another way. A categorial principle is 
a sort of purposive attitude taken in the interests 
of understanding and intelligibility with which we 
confront the given. It does not preclude any im- 
aginable or unimaginable content of experience in 
the future, but only precludes our interpreting it 
in a fashion contrary to our predetermined atti- 
tude or bent. 

If any be inclined to press the matter further, 
and raise the question how we can be positively 
assured that our minds will not alter in these fun- 
damental attitudes, I shall reply that we can not 
have any such final assurance. And it is not im- 
portant that we should. For a theory which re- 
fers the a priori to a transcendent, absolute, and 
universal mind, this question has its difficulties. 
But the theory here presented does not depend on 
the hypothesis of such a mind. It is compatible 
with the supposition that categorial modes of in- 
terpretation may be subject to gradual transition 
and even to fairly abrupt alteration. As will be 
pointed out in the next chapter, such alteration 
in categorial interpretations is a fact of social 
history but one which does not have the subver- 
sive results which might be imagined. There is no 
reason why it may not be a fact of the develop- 
ing mind of the individual as well. To be sure, the 


continuity of fundamental attitudes and purposes 
is the core of personality; the supposition that, 
without any rationale, these may become altered, 
is simply the supposition that a new and abnormal 
personality may replace our present one* This is 
admittedly possible, but it is not a contingency 
against which the theory of knowledge is supposed 
to provide* 





In experience, mind is confronted with the 
chaos of the given. In the interest of adaptation 
and control, it seeks to discover within or im- 
pose upon this chaos some kind of stable order, 
through which distinguishable items may become 
the signs of future possibilities. Those patterns 
of distinction and relationship which we thus seek 
to establish are our concepts. These must be de- 
termined in advance of the particular experience 
to which they apply in order that what is given 
may have meaning. Until the criteria of our in- 
terpretation have been fixed, no experience could 
be the sign of anything or even answer any ques- 
tion. Concepts thus represent what mind brings to 
experience. That truth which is a priori rises 
from the concept itself. This happens in two 
ways. In the first place, there is that kind of truth, 
exemplified most clearly by pure mathematics, 
which represents the elaboration of concepts in 
the abstract, without reference to any particular 
application to experience. Second, the concept in 
its application to the given exhibits the predeter- 
mined principles of interpretation, the criteria of 



our distinguishing and relating, of classification, 
and hence the criteria of reality of any sort. This 
is most clearly evident in the case of those basic 
concepts, determining major classes of the real, 
which may be called the categories, though in less 
important ways it holds true of concepts in gen- 

For both these ways in which the truth is fixed, 
independently of experience or in advance of it, 
it represents the explication or elaboration of the 
concept itself. Tlie a priori is not a material truth, 
delimiting or delineating tlie content of experience 
as such, but is definitive or analytic in its nature. 

The a priori as thus definitive or explicative, 
representing principles of order and criteria of 
the real, meets all the requirements which emerge 
from the discussion of the preceding chapter. 
Since it is a truth about our own interpretative 
attitude, it imposes no limitation upon the future 
possibilities of experience; that is a priori which 
we can maintain in the face of all experience, come 
what will. And although it represents the contri- 
bution of the mind itself to knowledge, it does not 
require that this mind be universal, absolute, or 
a reality of a higher order than the object of its 
knowledge. The a priori does not need to be con- 
ceived as the inscrutable legislation of a transcen- 
dent mind, the objects of which, being limited by 
its forms of intuition, are phenomenal only* Hence 
the distinction of the legislative mind as ultimate 


reality from its object which is not thus ultimate, 
falls away, and with it the difficulty of knowing the 
mind and of recognizing what is a priori as that 
which is determined by our own active attitude. 
The a priori is knowable simply through the re- 
flective and critical formulation of our own prin- 
ciples of classification and interpretation. Such 
legislation can be recognized as our own act be- 
cause the a priori principle which is definitive, and 
not a material truth of the content of experience, 
lias alternatives. It can be recognized as due to 
the mind itself by the ordinary criteria of respon- 
sibility in general that a different mode of act- 
ing is possible and makes a discoverable difference. 
Where there is no possibility of refraining from 
our act or acting otherwise, there can be no dis- 
coverable activity indeed, there is no act. As 
has been pointed out, if what is a priori sprang 
from a transcendent mind, acting in unalterable 
ways, it never could be known to be our own crea- 
tion or distinguished from those facts of life 
which are due to the nature of the independent 
real. What can be known to be a priori must meet 
the apparently contradictory requirements that it 
may be known in advance to hold good for all ex- 
perience and that it have alternatives. The prin- 
ciple of classification or interpretation meets these 
requirements, because the alternative to a defini- 
tion or a rule is not its falsity but merely its aban- 
donment in favor of some other. Thus the deter- 


mination of the a priori is in some sense like free 
choice and deliberate action* 

This meets also another difficulty which will al- 
ready have presented itself to the reader. If the a 
priori is something made by mind, mind may also 
alter it. There will be no assurance that what is a 
priori will remain fixed and absolute throughout 
the history of the race or for the developing indi- 
vidual. From the point of view here presented, 
this is no difficulty at all but the explanation of 
an interesting historical fact* The rationalist 
prejudice of an absolute human reason, universal 
to all men and to all time, has created an artifi- 
cially exalted and impossible conception of the 
categories as fixed and unalterable modes of mind. 
One result has been to limit the usefulness of the 
conception, so that what we could call, in ordinary 
parlance, "the categories of physics" or "the cate- 
gories of biology" would not serve as examples of 
4fi the categories" because it is obvious that the 
fundamental principles and concepts of any nat- 
ural science change progressively with its devel- 
opment. This, in turn, has served to obscure the 
large and important part played in science by 
that element of categorial order which cannot be 
determined by merely empirical fact but must be 
provided by the scientist himself in his setting of 
the problem and fixing the criteria by which the 
meaning of experimental findings is to be inter- 
preted. Thus the most impressive examples of hu- 


man knowledge have been too little drawn upon 
in discussions of epistemology. 

The assumption that our categories are fixed 
for all time by an original human endowment, is 
a superstition comparable to the belief of primi- 
tive peoples that the general features of their life 
and culture are immemorial and of supernatural 
origin. The grand divisions of our thought-world 
differ from those of our early ancestors as our 
modern machines differ from their primitive 
artifacts and our geographical and astronomical 
outlook from their world bounded by a distant 
mountain range or the pillars of Hercules and 
shut up under the bowl of the sky. Certain fun- 
damental categories are doubtless very ancient 
and permanent: thing and property , cause and 
effect, mind and body, and the relations of valid 
inference, doubtless have their counterparts wher- 
ever and whenever the human mind has existed. 
But even here, the supposition of complete iden- 
tity and continuity is at variance with facts which 
should be obvious. 

For all primitive peoples, for example, and for 
some who distinctly are not primitive, the prop- 
erties of a thing are not localized in time and 
space, as for us. Almost anything may be a talis- 
man or fetish, whose action takes place (without 
intermediaries) at a distance and in a time pos- 
terior to its destruction by fire or by being eaten. 
Things also have doubles, inscrutably operating 


in that other-world whose influence mysteriously 
interpenetrates the realm which we call "nature." 
Furthermore, the long-persistent problem in phys- 
ics of action at a distance increasingly comes back 
to haunt us and to unite with new problems of 
physical interpretation which threaten to drive 
us once more to dissipate the "material thing" 
throughout all time and space; to find its mani- 
festation and even its very being in a spatio-tem- 
poral spread of events indefinitely extended. 

That the present distinction of mind and body 
corresponds only roughly with that division in an- 
cient thought ; that body of inert matter and mind 
which does not occupy space are no older than the 
advent of that esoteric doctrine which dawned in 
Europe with the Greek mysteries and Christian- 
ity this can hardly escape us. This mode of dis- 
tinction contrasts with the tripartite division into 
body, mind, and spirit, and with the five-fold and 
Tfc-fold divisions of more easterly cultures. It is 
also obvious that the pressure of modern science 
in the field of biology and our present uneasiness 
about this twofold nature of the individual, augur 
some departure from the clarity of Cartesian 

The names of our categories may be very old 
and stable, but the concepts^ the modes of clas- 
sifying and interpreting which they represent, 
undergo progressive alteration with the advance 
of thought. 


Probably tliose modes of thought embodied in 
logic and in the forms of language are more fun- 
damental than others. And very likely what we 
recognize as explicit categories are always super- 
ficial as compared with more deep-lying forms 
which only the persistent and imaginative student 
can catch, in some vague and fleeting insight, be- 
cause they are so nearly the marrow of our being 
and so all-pervasive that they can hardly be 
phrased in significant expression. These go back 
to the point where mind is continuous with the ob- 
jective and indistinguishable from it. For we can 
know our own nature only in so far as we com- 
prehend or vaguely imagine what it would mean 
to be other than we are* We can recognize the 
presence of mind only where mind makes a dis- 
coverable difference. If we should think of mind 
as what the rationalists suppose superimposing 
on reality a rigid mask of form outside which 
mind itself could never catch a glimpse then this 
altogether universal and un-get-overable form 
could never become self-conscious. It would re- 
main in Fichte's phrase the "Great Thought 
which no man has ever thought." It would be not 
of mind but of the objective reality; it would be 
the Absolute which forever conditions but never 
can be known. But the idealistic rationalist can 
not eat his cake and have it too ; the mind which 
can be recognized as such is ipso facto finite and 
limited by discoverable bounds. That mind is thus 


continuous with the finally mysterious the is-ness 
of what is we must of course grant; in the con- 
templation of mind we contemplate one aspect of 
the Great Fact in the presence of wiiich all ex- 
plicit thought is silenced. But the categories are 
not the form of that which, having no alternative 
and no bounds, is formless. They are the explicit 
bounds of that which, if it transcend them, must 
fall into some other category. They are di- 
visions within the comprehensible in general, but 
not the shape of comprehensibility itself* 

It will be well to make clear that the concep- 
tion here presented does not imply that because 
the a priori is something made by mind and capa- 
ble of alteration, it is therefore arbitrary in the 
sense of being capriciously determined. That it is 
not, and cannot be, determined by the given, 
does not imply that it answers to no criteria what- 
ever. That type of a priori truth which pure 
mathematics illustrates that is, the elaboration 
of concepts in abstraction from all questions of 
particular applications answers only to the cri- 
teria of self -consistency. Just to the extent that 
the development of such a purely analytic sys- 
tem is withdrawn from every consideration of use- 
ful application, its truth is simply truth to the 
original meanings embodied in its basic concepts. 
But when concepts are intended to be applied in 
experience, and a priori principles are to deter- 
mine modes of classification and interpretation. 


the case is different. Here mind is still uncom- 
pelled by any possible content of experience. But 
knowledge has a practical business to perform, the 
interests of action which it seeks to serve. The 
mode of the mind's activity answers to our need 
to understand, in the face of an experience al- 
ways more or less baffling, and of our need to 
control* There is also another factor which helps 
to determine what modes of attempted compre- 
hension will be most easily and most widely use- 
ful. While that absolute human reason which the 
rationalist supposes to be completely and univer- 
sally possessed by every human is a myth, never- 
theless man, being a species of animal, has char- 
acteristics which mark him as such, and some of 
these at least are reflected in the bent of human 
thought. Some modes of thought are simpler and 
come more naturally to us than others which still 
are possible and which might, indeed, be called 
upon if an enlarged experience should sufficiently 
alter our problems just as some modes of bodily 
translation are more easy and natural, though 
these may be somewhat altered when the environ- 
ment includes a sufficient number of automobiles 
and airplanes. Moreover, the fundamental like- 
ness in our modes of thought, which represents 
whatever community of nature marks our original 
mental endowment, is continually enhanced by the 
fact that the needs of individual humans are 
mostly served by cooperation with others. "The 


human mind" is distinctly a social product, and 
our categories will reflect that fact. 

In brief, while the a priori is dictated neither 
by what is presented in experience nor by any 
transcendent and eternal factor of human nature, 
it still answers to criteria of the general type 
which may be termed pragmatic. The human ani- 
mal with his needs and interests confronts an ex- 
perience in which these must be satisfied, if at all. 
Both the general character of the experience and 
the nature of the animal will be reflected in the 
mode of behavior which marks this attempt to 
realize his ends. This will be true of the cate- 
gories of his thinking as in other things. And 
here, as elsewhere, the result will be reached by a 
process in which attitudes tentatively assumed, 
disappointment in the ends to be realized, and 
consequent alteration of behavior will play their 

Confirmation of this conception of the a priori 
could only come from comprehensive and detailed 
examination of at least the major categories of 
thought and the underlying principles of com- 
mon-sense and scientific explanation. Such a task 
cannot be undertaken here; at most only a few 
illustrations can be offered with the hope that they 
are typical. 

The paradigm of the a priori in general is the 
definition. It has always been clear that the sim- 
plest and most obvious case of truth, which can 


be known in advance of experience is the explica- 
tive proposition and those consequences of defini- 
tion which can be derived by purely logical analy- 
sis. These are necessarily true, true under all 
possible circumstances, because definition is legis- 
lative. Not only is the meaning assigned to words 
more or less a matter of choice that considera- 
tion is relatively trivial but the manner in which 
the precise classifications which definition embod- 
ies shall be affected, is something not dictated by 
experience. If experience were other than it is, the 
definition and its corresponding classification 
might be inconvenient, useless, or fantastic, but 
it could not be false. Mind makes classifications 
and determiiies meanings; in so doing, it creates 
that truth without which there could be no other 

Traditionally propositions which have been 
recognized as analytic have often not been classed 
with the a priori; they have been regarded as too 
unimportant; sometimes they have even been re- 
pudiated as not truth at all but merely verbal 
statements- The main reasons for this cavalier at- 
titude have been two; in the first place, it has 
been, overlooked that the real itself is a matter of 
definition and that the dichotomy of real and un- 
real is that first and basic classification which the 
mind confronted with experience must make. And 
second, the powerful sweep and consequence of 
purely logical analysis has not been understood. 


The clearest example of this power of analysis 
is to be found, of course, in mathematics. The his- 
torical importance of mathematics as a paradigm 
of a priori truth needs no emphasis- Almost one 
may say that traditional conceptions of the a 
priori are the historical shadow of Euclidean 
geometry. But in mathematics much water has 
gone under the bridge since the time of Kant, and 
in the light of the changes which have come about, 
these traditional conceptions are proved totally 
impossible. The course of this development will be 
familiar to the reader; only the outstanding fea- 
tures of it need be mentioned. 

Though there are anticipations of current 
mathematical conceptions as far back as Plato, 
the movement which led to their present accep- 
tance dates principally from the discovery of the 
non-Euclidean geometries. In developing these 
systems, it was obviously impossible to depend on 
intuitions of space, either pure or empirical. If 
Euclid is true of our space, then no one of these 
geometries can be; and if Euclid is not true and 
certain, then the main ground of the supposi- 
tion that we can rely on intuitions of the spatial 
is discredited.* Hence in developing the non- 
Euclidean systems, all constructions such as help- 
ing-lines, and any step in proof which should de- 

*Euclid and a non-Euclidean system cannot both be true of space 
while corresponding denotations of terms are maintained. The dis- 
covery that they may both become true of space -with systematic 
difference in the denotation of terms played its part in the logic of 
modern geometry. 


pend not upon pure logic but upon the character 
of space must be dispensed with. If a step in proof 
cannot be taken by rigorous logic alone, it can- 
not be taken at all. When it was found thus pos- 
sible to develop the non-Euclidean systems with- 
out appeal to any extra-logical aids, a similar re- 
vision of Euclid was carried out, eliminating all 
explicit or implicit reliance upon constructions, 
superpositions or other appeal to spatial intui- 
tion. This new method, together with certain in- 
dicated generalizations, constituted the so-called 
4t modern geometry.' 5 

Next it was demonstrated that not only geome- 
try but other branches as well can be developed by 
the deductive method, from a relatively few as- 
sumptions, and likewise without reliance upon 
empirical data. As a result all pure mathematics 
is found to be abstract, in the sense of being in- 
dependent of any particular application. Because 
if all the theorems follow logically from the defini- 
tions and postulates, then we can alter at will what 
we let the terms, such as "point" and "line," de- 
note without in the least disturbing any step in 
the proofs. Whatever "point" and "line" may 
mean, given these assumptions about them, these 
consequences the rest of the system must also 
hold of them, since the theorems follow from the 
assumptions by rigorous and purely logical de- 

The question of the truth of the mathematical 


S3 r stem in application was thus complete!}- sepa- 
rated from its mathematical or logical integrity. 
Still further changes went along with this. The 
"truth'* of initial assumptions lost all meaning in 
any other sense than their exhibition of certain 
patterns of logical relationship to be adhered to 
throughout. The distinguishing assumptions of a 
non-Euclidean geometry, for example, so far 
from being self-evident, were supposedly mere ar- 
bitrary falsehoods with respect to their most ob- 
vious empirical denotation. The term "axiom" was 
replaced by "postulate" or "primitive proposi- 
tion.' 5 In the interest of logical simplicity alterna- 
tive sets of assumptions which would give the 
same system of propositions were investigated. 
What should be initially assumed and what proved 
became a question merely of such logical simplic- 
ity. It became customary to speak of the truth of 
mathematics as hypothetical or to say that what 
mathematics asserts is only the relation of impli- 
cation between postulates and theorems. It is truth 
about certain patterns of logical relationship es- 
tablished by initial definition or postulate. 

Further, it became clear that the distinction be- 
tween those assumptions of the form called "defini- 
tions" and those termed "postulates" was relatively 
arbitrary and unimportant. Logically it mates 
little difference except for simplicity of proced- 
ure, how far the order of a system is set up by 
propositions in which "is" means logical equiva- 


lence and how far by those in which it means only 
the one-way implication of concepts or subsump- 
tion of classes. Since the content of the concepts 
of pure mathematics is simply that order to which 
they give rise, that manner of development in 
which essential relationships are exhibited as the 
definitive meaning of the concepts is truest to the 
nature of the subject. 

The completion of this last refinement of math- 
ematical method was made by Whitehead and 
Russell in "Principia Mathematical* It was here 
proved that the initial assumptions of mathe- 

"This development is still too recent for the exact bearing of its 
consequences to be clear in all respects. Points concerning which there 
may be some doubt are as follows: 

(1) Can this method, without alteration, be completely carried out 
In geometrical branches? Some concepts needed in geometry are 
dealt with in volume HI, but volume IV, which was to complete this 
subject, has not appeared. 

(2) There are certain assumptions, such as the "axion of infinity" 
which appear as hypotheses to some theorems which, apparently, re- 
quire diem. It is not known whether this procedure is entirely avoid- 
able without abandoning certain classes of theorems which have then? 
place in usual mathematical developments. 

(3) Consistency and independence are of the essence of mathe- 
matics, since^they concern its logical integrity, apart from all ques- 
tions of application. It is possible to think, in view of usual mathe- 
matical procedures, that tests of consistency and independence in- 
volve implicit appeal to intuition or to applications. By the method 
of "Principia*' it is not clear how it is possible to deal with such 
problems. Independence proofs have been applied to logistic systems 
by N. Bernays and others, but only by a reversion to those familiar 
devices the logic of which is precisely the point in question. Professor 
H. M. Sheffer has offered a general method for testing consistency 
and independence without reference to any possible application (See 
his "General Theory of Notational Relativity," privately circulated). 
The method of dealing with usual branches in this way must wait 
upon his further publication. 

^ W^It is questionable whether logistic methods employed in "Prin- 
cipxa" are in all respects acceptable as "proof" and whether they do 
not, in certain ways, import illicit assumptions of "existence." But 
if there are defects of this sort, it is highly probable that they ate 


matics can all be dispensed with, except the defini- 
tions. The truths of mathematics follow merely 
from definitions which exhibit the meaning of its 
concepts, by purely logical deduction. Judgment 
of such mathematical truth is, thus, completely 
and exclusively analytic; no synthetic judgment, 
a priori or otherwise, is requisite to knowledge of 
pure mathematics. The content of the subject con- 
sists entirely of the rigorous logical analysis of 
abstract concepts, in entire independence of all 
data of sense or modes of intuition. The defini- 
tions which embody these concepts are not re- 
quired to be true in any other sense than that 
they should be precise and clear ; the formulation 
of them represents an act of mind which is legis- 
lative or creative and in some sense arbitrary; it 
answers to no criteria save self -consistency and 
adequacy to whatever purposes the elaboration of 
the system itself may be supposed to satisfy* It 
may still be true that "concepts without precepts 
are empty,' 5 but it must be granted that there is 
a kind of knowledge of "empty" concepts. Or at 
least such admission can be avoided only by a re- 
striction of the term "knowledge" to exclude pure 
mathematics and logic. The importance of such 
a priori analytic knowledge is witnessed by the 
basic character of these subjects for all other sci- 

Pure mathematics stands between logic on the 
one side and the empirical application of mathe- 


matics on the other. Logic is in some respects the 
illustration par excellence of the a priori, since its 
laws are the most completely general of any. The 
laws of logic cannot be proved unless they should 
first be taken for granted as the principles of 
their own demonstration. They make explicit the 
basic principles of all interpretation and of our 
general modes of classification. And they impose 
no limitation upon the content of experience. 
Sometimes we are asked to tremble before the 
specter of the "alogical" in order that we may 
thereafter rejoice that we are saved from this by 
the dependence of reality upon mind. But the 
"alogical" is pure bogey, a word without a mean- 
ing. What kind of experience could defy the prin- 
ciple that everything must either be or not be, 
that nothing can both be and not be, or that if 
X is Y and Y is Z, then X is Z? If anything im- 
aginable or unimaginable could violate such, laws, 
then the ever-present fact of change would do it 
every day* The laws of logic are purely formal; 
they forbid nothing but what concerns the use of 
terms and the corresponding modes of classifica- 
tion and analysis. The law of contradiction tells 
us that nothing can be both white and not white, 
but it does not and can not tell us whether black 
is not white or soft or square is not white. To dis- 
cover what contradicts what we must turn to more 
particular considerations. Similarly the law of the 
excluded middle formulates our decision that 


whatever is not designated by a certain term shall 
be designated by its negative. It declares our pur- 
pose to make, for every name, a complete dichot- 
omy of experience, instead as we might choose 
of classifying on the basis of a tripartite di- 
vision into opposites and a middle ground between 
the two. Our rejection of such tripartite division 
represents only our penchant for simplicity and 
similar considerations* 

Further laws of logic are of like significance. 
They are principles of procedure, the parliamen- 
tary rules of intelligent thought and action. Such 
laws are independent of the given because they 
impose no limitations whatever upon it. They are 
legislative because they are addressed to ourselves 
because definition, classification, and inference 
represent no operation in the world of things, but 
only our categorial attitudes of mind. 

Furthermore, the ultimate criteria of the laws 
of logic are pragmatic. Indeed, how could they 
be anything else? The truth of logic is not ma- 
terial truth but a truth about the modes of self- 
consistency. Since this is so, logic must be the test 
of its own consistency, and hence of its own truth, 
as well as the test of the consistency of everything 
else. But if logic tests its own truth, then what 
can be the test of truth in a genuine issue of logic, 
which is not a question of mere inadvertance on 
one side or the other? Those who suppose that 
there is a logic which everyone would agree to if 


he understood it and understood himself, are more 
optimistic than those versed in the history of logi- 
cal discussion have a right to be. The fact is that, 
as was pointed out in the preceding chapter, there 
are several logics, markedly different, each self- 
consistent in its own terms and such, that who- 
ever, using it, avoids false premises, will never 
reach a false conclusion, Mr. Russell, for exam- 
ple, bases his logic on an implication relation such 
that if twenty sentences be cut from a newspaper 
and put in a hat, and then two of these be drawn 
at random, one of them will certainly imply the 
other, and it is an even chance that the implica- 
tion will be mutual. Yet upon a foundation so re- 
mote from ordinary modes of inference the whole 
structure of "Principia Mathematica" is built. 
This logic is utterly self -consistent and valid in 
its own terms. There are others even more strange 
of which the same may be said.* Genuine issues 
of logic are those which stand above such ques- 
tions of the merely self -critical integrity of the 
logical system. There are such issues, and these 
cannot be determined nay, cannot even be ar- 
gued except on pragmatic grounds of human 
bent and intellectual convenience. That we have 
been blind to this fact, and that much good paper 
and ink has been wasted by logicians who have 
tried to argue on some other grounds what are 
only questions of convenience or of value, itself 

*See p. 209, footnote. 


reflects traditional errors in the conception of the 
a priori. 

Pure mathematics and logic exemplify that 
type of the a priori which have the highest de- 
gree of abstraction from experience whose con- 
cepts are so general that we may call them 
"empty." Concerning these, there may be a ques- 
tion whether there will not be issues of an entirely 
different sort when we attempt to apply them in 
experience. One may say, for example, that when 
geometry becomes abstract and freed from all 
necessary reference to our intuitions of the spa- 
tial, the question of the truth about space be- 
comes an entirely separate one, and one with re- 
spect to which there must be reference to forms 
of intuition or something of the sort, or there will 
be nothing which is determinable a priori at all. 
Similarly one may say that if arithmetic as a 
purely abstract deductive system has no necessary 
reference to the character of countable objects, 
then its a priori truth is of no value for the an- 
ticipation of the behavior of concrete things. This 
will be true, of course, and of importance. If 
there should be a priori truth only with respect to 
concepts in utter abstraction from experience, and 
if this a priori character were to vanish when 
these concepts are given a concrete denotation, 
then the significance of the a priori for the nat- 
ural sciences and for common practice would be 
largely, if not completely, lost. 


But there is an a priori truth of concepts which 
have concrete denotation. Let us consider the ex- 
ample of arithmetic. Arithmetic depends In toto 
upon the operation of counting or correlating, a 
procedure which can be carried out in any world 
containing identifiable things, regardless of the 
further characters of experience. Mill challenged 
this a priori character of arithmetic. He asked us 
to suppose a demon sufficiently powerful and 
maleficent so that every time two things were 
brought together with two other things, this 
demon should always introduce a fifth. The con- 
clusion which he supposed to follow is that under 
such circumstances 2 -f 2 = 5 would be a univer- 
sal law of arithmetic. But Mill was quite mis- 
taken. In such a world we should be obliged to 
become a little clearer than is usual about the 
distinction between arithmetic and physics, that 
is all. If two black marbles were put in the same 
box with two white ones, the demon could take his 
choice of colors, but it would be evident that there 
were more black marbles or more white ones than 
were put in. The same would be true of all ob- 
jects in any wise identifiable. We should simply 
find ourselves in the presence of an extraordinary 
physical law, which we should recognize as uni- 
versal in our world, that whenever two things were 
brought into proximity with two others, an addi- 
tional and similar thing was always created by 
the process. Mill's world would be physically most 


extraordinary. The world's work would be enor- 
mously facilitated if hats or locomotives or tons 
of coal could be thus multiplied by anyone pos- 
sessed originally of two pairs. But the laws of 
mathematics would not be affected- It is because 
this is true that arithmetic is a priori. Its laws 
prevent nothing; they are compatible with any- 
thing which happens or could conceivably happen 
in nature. They are true in any possible world. 
Mathematical addition is not a physical trans- 
formation. The only bringing together it implies 
is in the mind; if translation in general affected 
numerical alteration, we should always count 
things in $itu y but we should count and add as 
usual. Physical changes which result in an in- 
crease or decrease of the countable things involved 
are matters of everyday occurrence. Such physi- 
cal processes present us with phenomena in which 
the purely mathematical has to be separated out 
by analysis. It is because we shall always separate 
out that part of the phenomenon not in conf ormi- 
ity with arithmetic and designate it by some other 
category physical change, chemical reaction, op- 
tical illusion that arithmetic is a priori. Its laws 
constitute criteria of our categorial classification 
and interpretation. As this example serves to 
illustrate, such categorial interpretation of the 
concrete and empirical throws out of court what- 
ever would otherwise violate the a priori princi- 
ples which embody the category, but it does not 


thereby legislate anything phenomenal out of ex- 

Perhaps, however, we have gone too far. Mill's 
illustration is of an alteration of experience in 
general which is too simple and too poorly car- 
ried out to make it plausible that our categorial 
interpretation would be different in such a world. 
But if translation in general affected numerical 
alteration then an entirely different mode of cate- 
gorial interpretation might better serve the pur- 
poses. Our present categories would not could 
not be prohibited but other modes might more 
simply reduce the phenomenal to order and facili- 
tate control. Or in such a world, arithmetic might 
be confined to mental phenomena since these 
would be exempt from the effects of change of 
place and numerical principles would be laws of 
psychology. If we were jelly-fish in a liquid world, 
we should probably not add at all, because the 
useful purposes served by such conceptions would 
be so slight. Still if some super-jelly-fish should 
invent arithmetic by a jeu d'esprit (as Hamilton 
invented quaternions) he would find nothing in 
any possible experience to controvert it, and he 
might with some profit apply it to his own dis- 
tinct ideas. 

The ideal illustration of the a priori in applied 
geometry would be a consideration of physical 
relativity, showing how geometrical truth may 
turn upon the place where the dividing line is 


drawn between the properties of space and those 
of matter. Applied geometrical principles are a 
priori true of all space-filling things. But this 
a priori truth has its pragmatic aspect since there 
are these alternatives about the manner in which 
the category of "the spatial 55 shall be bounded. 
But to carry out this illustration in detail would 
be beyond my competence,* 

Incidentally it may be pointed out that the 
ideas which have gained currency through rela- 
tivity-theory make clear the nature of Kant's mis- 
takes in the supposition of a limited form of 
spatial intuition. For Kant, the spatial or geomet- 
rical has to do with relations of the simultaneous ; 
the shape of a triangle, for example, is something 
instantaneously imaginable. But for celestial tri- 
angles, such instantaneous intuition has no mean- 
ing; what exists or happens at a distance is not 
directly verifiable here and now; a passage of 
something through time, as well as space, is inex- 
tricably bound up with the determintion of the 
distant fact* Hence the imagination of a "curved- 
space*' need not mean something like flattening 
out a hemisphere without disturbing the relations 
of great circles on its surface. It means only im- 
agining certain uniform sequences to characterize 
our experience of the spatial under certain condi- 
tions. The "unimaginable 55 character of curved- 

*The reader may also be referred to Poincare's " Science and Hypo- 
thesis," the section on " Space and Geometry/* 


space In the sense that we cannot visualize a non- 
Euclidean triangle on the blackboard, has noth- 
ing to do with the matter. Triangles on different 
scales have different "shapes" in non-Euclidean 
space, and triangles big enough to "verify" the 
nature of space are too big to "imagine." Our an- 
cestors who believed the earth was flat could cer- 
tainly "imagine" a non-Eucidean space, in the 
only sense which is required, since the geometry 
of the earth's surface is (in an obvious sense) 

The a priori element in natural sciences goes 
much deeper than might be supposed. All order 
of sufficient importance to be worthy of the name 
of law depends eventually upon some ordering by 
mind. Without initial principles by which we guide 
our attack upon the welter of experience, it would 
remain forever chaotic and refractory. In every 
science there are fundamental laws wiiicli are a 
priori because they formulate just such definitive 
concepts or categorial tests by which alone inves- 
tigation becomes possible. 

A good example of this is to be found in Ein- 
stein's little book "Relativity."* The question un- 
der discussion is the criteria of simultaneity for 
events at a distance. Suppose the lightning strikes 
a railroad track at two places, A and B. How 
shall we tell whether these events happen at the 
same time? "We . . . require a definition of simul- 
taneity such that this definition supplies us with 

*Pp. 26-28: italics axe the author's. 


a method by which ... we can decide whether or 
not the lightning strokes occurred simultaneously- 
As long as this requirement is not satisfied, I allow 
myself to be deceived as a physicist (and of course 
the same applies if I am not a physicist) when I 
imagine that I am able to attach a meaning to the 
statement of simultaneity* . . . 

"After thinking the matter over for some time 
3 T ou then offer the following suggestion with which 
to test simultaneity. By measuring along the rails, 
the connecting line A B should be measured up 
and an observer placed at the mid-point M of the 
distance A B. This observer should be supplied 
with an arrangement (e. g., two mirrors inclined 
at 90) which allows him visually to observe both 
places A and B at the same time. If the observer 
perceives the two flashes at the same time, they are 

"I am very pleased with this suggestion, but 
for all that I cannot regard the matter as quite 
settled, because I feel constrained to raise the fol- 
lowing objection: Tour definition would certainly 
be right, if I only knew that the light by means 
of which the observer at M perceives the lighting 
flashes travels along the length A-M with the 
same velocity as along the length B M . But an 
examination of this supposition would only be pos- 
sible if we already had at our disposal the means 
of measuring time. It would thus appear as though 
we were moving here in a logical circle** 

"After further consideration you cast a some- 


what disdainful glance at me and rightly so 
and you declare: *I maintain my previous defini- 
tion nevertheless because in reality it assumes 
nothing whatever about light. There is only one 
demand to be made of the definition of simultane- 
ity, namely, that in every real case it must supply 
us with an empirical decision as to whether or not 
the conception which has to be defined is fulfilled. 
That light requires the same time to traverse the 
path AM. as for the path BM is in reality 
neither a supposition nor a hypothesis about the 
physical nature of light, but a stipulation which 
I can make of my own free-will in order to arrive 
at a definition of simultaneity.' . . . We are thus 
led to a definition of *time ? in physics." 

As this example well illustrates, we cannot even 
ask the questions which discovered law would an- 
swer until we have first by a priori stipulation 
formulated definitive criteria. Such concepts are 
not verbal definitions nor classifications merely; 
they are themselves laws which prescribe a certain 
behavior to whatever is thus named. Such defini- 
tive laws are a priori ; only so can we enter upon 
the investigation by which further laws are 
sought. Yet it should also be pointed out that such 
a priori laws are subject to abandonment if the 
structure which is built upon them does not suc- 
ceed in simplifying our interpretation of phe-* 
nomena. If, in the illustration given, the relation 
"simultaneous with*" as defined, should not prove 


transitive if event A should prove simultaneous 
with B and B with C, but not A with C this defi- 
nition would certainly be rejected. 

Indeed all definitions and all concepts exercise 
this function of prescribing fundamental law to 
whatever they denote, because everything which 
has a name is to be identified with certainty only 
over some stretch of time. The definition provides 
criteria of the thing defined which, in application, 
become necessary or essential laws of its behavior. 
This is especially evident in the case of scientific 
definitions because the "things" of science are of 
a deep-lying sort, representing uniformities of be- 
havior of a high order.* If definition is unsuccess- 
ful, as early scientific definitions mostly have been, 
it is because the classification thus set up corre- 
sponds with no natural cleavage and does not cor- 
relate with sufficiently important uniformities of 
behavior. Early attempts to reduce phenomena to 
law are based upon the "things" of common-sense 
which represent classification according to proper- 
ties which are relatively easy of direct observation 
and impressive to the senses. When such attempts 
fail, it is largely because of this superficiality of 
initial classification. The alchemist's definitions of 
the elements, for example, are the clue to his in- 
different success ; the definitive properties pick out 
amorphous groups which have little significance 
of further uniformity. Not until such crucial 

*See Appendix A. 


properties as combining weights become the basis 
of classification is it possible to arrive at satisfac- 
tory laws of chemistry. The earlier definitions can 
not be said to have been false; they were merely 
useless, or insufficient to the purposes in hand. A 
large part of the scientific search is, thus, for 
things wortJi naming. 

We have reached a point today where we un- 
derstand that the typical procedure of science is 
neither deduction from what is self-evident nor 
relatively certain nor direct generalization from 
experience. If any one method is more charac- 
teristic of science than another it is that of hy- 
pothesis and verification. But we seem still to over- 
look the fact that the terms in which "hypothesis 
and law are framed themselves represent a scien- 
tific achievement. We still suffer from the delusion 
that fixed and eternal categories of human thought 
on the one side are confronted with equally fixed 
and given "things*' on the other. Is it not obvious, 
to dispassionate observation, that scientific cate- 
gories and classifications are subject to progres- 
sive modification or even abrupt alteration, and 
that these have a directive and controlling influ- 
ence upon the other phases of scientific research? 
And here too, as in hypothesis and verification, the 
development takes place, not by logical derivation 
from antecedent principles nor by direct formula- 
tion of empirical content, but by the hazarding of 
something by the mind and Its retention or re- 


pudiation according to the success or non-success 
of what is based upon it. The test of success here 
is not, however, simple conformity with experi- 
ences as in the testing of hypothesis, but is the 
achievement of intelligible order amongst the phe- 
nomena in question, and responds also to such 
criteria as intellectual simplicity, economy, and 
comprehensiveness of principle. 

The reader will perhaps feel that, in so far as 
this is true, what is here represented as a priori 
is nothing of the sort but is merely something that 
we learn from experience. But if so, I hope that 
he will reread the illustration from Einstein, with 
due regard for the logic of it. However much the 
give and take between the purposes of science and 
discovered fact may contribute to alter the pro- 
cedure by which those aims are sought, and may 
induce new basic principles and categories, still 
the naming, classifying, defining activity is at 
each step prior to the investigation. We cannot 
even interrogate experience without a network of 
categories and definitive concepts. Until our mean- 
ings are definite and our classifications are fixed, 
experience cannot conceivably determine anything. 
We must first be in possession of criteria which 
tell us what experience would answer what ques- 
tions, and how, before observation or experiment 
could tell us anything. 

The uniformities which science seeks are of a 
high order and represent a further reach of those 


same purposes exhibited on a humbler scale by the 
uniformities of common sense comprehension. The 
categorizing and classifying activity of thought 
is, thus, more deliberate and self-conscious in the 
case of science and, by comparison, easier to ob- 
serve than in the case of common sense. Because 
scientific categories are in some part built upon 
more basic distinction, the functioning of them as 
criteria of the real is less frequent and perhaps 
less important. It is, nevertheless, sometimes to be 
observed in the interplay between new principles 
of scientific interpretation and residual phenomena 
which are unexplained or reports of observation 
which are regarded as possibly involving error. 
If Rontgen had been unable to repeat the experi- 
ence in which he first saw the bones of his hand, 
that perception would have been discredited. Be- 
cause no one but their discoverer could see the 
"N-rays" with any assurance, and it was found 
that he saw them when the refracting prism was 
removed as well as with it, they were discarded as 
illusory. The phenomena of hypnotism remained 
for a long time in the limbo of the dubious, and 
even today offer a difficult problem of separation 
of veridical phenomenon from illusion, self-decep- 
tion and the like. The phenomena of "dual per- 
sonality" would challenge a very fundamental 
category if only there were not so much doubt 
about what it is that is "dual** and whether it is 
genuinely dual. To choose a different sort of ex- 


ample, when the eclipse-photographs which fig- 
ured in the discussion of relativity were examined* 
the question was raised whether the star-displace- 
ments as measured on them represented simply 
the bending of light-rays or were due in part to 
halation of the sensitive film. Thus, for a moment 
at least, so fundamental a problem as that of 
abandoning the categories of independent space 
and time was intertwined with the question whether 
the position of dots on a photographic plate rep- 
resented authentic star-photographs or was due to 
something which took place inside the camera. 

What needs to be observed here is at once the 
continuity of scientific problems of a high order 
with the apparently simple and fundamental cri- 
teria of the real, and the fact that such decisions 
of reality or unreality are themselves interpreta- 
tions involving principles of the same order as 
scientific law. They are such as forbid, for ex- 
ample, the non-biological transformations and 
non-physical successions which occur in dreams. 
A mouse which disappears where there is no hole, 
is no real mouse ; a landscape which recedes as we 
approach, is but illusion. The reality of an object 
of a particular sort is determined by a certain 
uniformity of its behavior in experience. The for- 
mulation of this uniformity is of the type of natu- 
ral law. So far, such laws are a priori f or this 
particular sort of thing; the experience which fails 
to conform to the law is repudiated as non-veridi- 


This situation is most paradoxical; principles 
of the order of natural law are reached by some 
generalization from experience that is, from ve- 
ridical experience; there are no generalizations 
whatever to be had from the unsorted experience 
of real and unreal both. But what experience is 
veridical, is determined by the criterion of law. 
Which is first, then; does the content of experi- 
ence validate the law or does the law validate the 
experience in attesting its veridical character? 
The answer is that the law is first precisely so long 
as and so far as we are prepared to maintain it 
as criterion of the real. But the "reality" which 
is in question is likely to be of a highly specific 
sort. The authentic photograph of stars and the 
picture affected by halation, for example, are both 
real, both physically real, both photographs even, 
in a certain sense. What is an "authentic photo- 
graph;" has to be very precisely defined in the 
case in point and is so defined as to exclude the 
effect of light reflection from the back of the 
glass of the photographic plate. The manner of 
this definition or classification obviously will re- 
quire a correlation of photograph and thing 
photographed of the type set forth in certain 
physical laws. In the particular case, failure to 
exhibit such lawful relationship condemns the phe- 
nomenon as not authentic or not jeal that is, not 
really this very specific sort of thing. 

Thus all concepts, and not simply those we 


should call "categories,'* function as criteria of 
reality. Every criterion of classification is cri- 
terion of reality of some particular sort. There is 
no such thing as reality in general; to be real, a 
thing must be a particular sort of real. Further- 
more, what is a priori criterion of reality in one 
connection may be merely empirical law in some 
other for example, the law correlating photo- 
graph and thing photographed, or the law of the 
behavior of solid bodies in translation which con- 
demns the mouse that disappears without a hole, 
or the laws of perspective which exclude a land- 
scape which recedes as we approach it. The de- 
termination of reality, ike classification of phe- 
nomena, and the discovery of law, all grow up 
together. I will not repeat what has already been 
said so often about the logical priority of criteria ; 
but it should be observed that this is entirely com- 
patible with the shift of categories and classifica- 
tions with the widening of human experience. If 
the criteria of the real are a priori, that is not to 
say that no conceivable character of experience 
would lead to alteration of them. 

For example, spirits cannot be photographed. 
But if photographs of spiritistic phenomena, 
taken under properly guarded conditions, should 
become sufficiently frequent, this a priori dictum 
would be called in question. What we should do 
would be to redefine our terms. Whether "spook" 
was spirit or matter, whether the definition of 


"spirit 55 or of "matter" should be changed; all 
this would constitute one interrelated problem. 

What would prove to you that the relative mo- 
tion of a body effected a foreshortening of it and 
altered its mass? If the answer is "no conceivable 
experience'* and you are able to formulate a defi- 
nition of "mass," a conception of motion, and of 
ideally exact measurement in terms that do not 
conflict with one another and with your other 
physical conceptions, there is no possible ground 
on which you could be proven wrong. To a mind 
sufficiently resolute for an independent space and 
time, no possible experience could prove the prin- 
ciples of relativity. The question, "How long shall 
we persist in holding to our previous categories, 
when confronted with star-photographs and the 
displacement of spectrum-lines (or with spiritistic 
phenomena or evidence of telepathy)?" is one 
which has no general answer. A stubborn con- 
servatism can be proved unreasonable only on the 
pragmatic ground that another method of cate- 
gorial analysis more successfully reduces all ex- 
perience of the type in question to order and law. 
Confronting such a problem, we should reopen to- 
gether the question of definition or classification, 
of criteria for this sort of real, and of natural 
law. And the solution of one of these would prob- 
ably mean the solution of all. Nothing could force 
a redefinition of "spirit" or of "matter." A suf- 
ficiently fundamental relation to human bent or 


to human interests would guarantee continuance 
unaltered even in the face of unintelligible and 
baffling experience. And no equipment of cate- 
gories and concepts which the mind is likely to 
achieve will enable us to understand experience 
completely and in every respect. In such problems, 
the mind finds itself uncompelled save by its own 
purposes and needs. What is fixed datum and must 
be conformed to, is only that welter of the given 
in which not even the distinction of real and un- 
real is yet made. The rest is completely and ex- 
clusively our problem of interpretation, I may 
categorize experience as I will; but what cate- 
gorial distinctions will best serve my interests and 
objectify my own intelligence? What the mixed 
and troubled experience will be that is beyond 
me. But what I shall do with it that is my own 
question, when the character of experience is be* 
fore me. I am coerced only by my own need to 

It would indeed be inappropriate to charac- 
terize as a priori a law which we are prepared to 
alter in the light of further experience even 
though in an isolated case we should discard as 
non-veridical any experience which failed to con- 
form. But the crux of the matter lies in this; be- 
yond such principles as those of logic and pure 
mathematics whose permanent stability seems at- 
tested, there must be further and more particular 
criteria of the real prior to any investigation of 


nature. Such definitions, fundamental principles 
and criteria the mind itself must supply before 
experience can even begin to be intelligible. These 
represent more or less deep-lying attitudes, which 
the human mind has taken in the light of its total 
experience up to date. But a newer and wider 
experience may bring about some alteration of 
these attitudes even though by themselves they 
dictate nothing as to the content of experience, 
and no experience can conceivably prove them in- 

It is the a priori element in knowledge which is 
thus pragmatic, not the empirical. The pragma- 
tists generally have neglected to make the sepa- 
ration of concept and immediacy, with the result 
that they seem to put all truth at once at the 
mercy of experience and within the power of hu- 
man decision or in a relation of dependence upon 
the human mind. But this would be an attempt to 
have it both ways. The sense in which facts are 
brute and given cannot be the sense in which the 
truth about them is made by mind or alterable to 
human needs. To be sure, this a priori element in 
knowledge runs very deep ; it is present whenever 
there is classification, interpretation, or the dis- 
tinction of real from unreal which means that 
it is present in all knowledge. So I suppose it must 
be admitted, in the last analysis, that there can 
be no more fundamental ground than the prag- 
matic for a truth of any sort. Nothing not even 


direct perception can force the abandonment of 
an interpretive attitude, nor indeed should move 
us to such abandonment (since illusion or mistake 
is always possible) except some demand or pur- 
pose of the mind itself. But certain important 
ends, such as intellectual consistency and econ- 
omy, completeness of comprehension, and simplic- 
ity of interpretation, occupy a place so much 
higher, for the long-run satisfaction of our needs 
in general, that they rightfully take precedence 
over any purpose which is merely personal or 
transitory. In the popular mind especially, prag- 
matism too often seems to connote the validity of 
rather superficial and capricious attitudes for 
instance, the justification of belief from no deeper 
ground than personal desire. It is this insufficient 
regard for intellectual integrity, this tendency 
to trench upon high-plane purposes from low- 
plane motives which marks the kind of "pragma- 
tism" which is to be eschewed. We must all be 
pragmatists, but pragmatists in Ihe end, not in 
the beginning. 

In another respect also, there is a connotation 
of "pragmatism" which more or less prevails, 
which is inapplicable to the theory here presented. 
Concepts and principles of interpretation are sub- 
ject to historical alteration and in terms of them 
there may be "new truth." But the situation in 
which this happens needs analysis. It does not 
mean the possibility of new truth in any sense in 


which new truth can genuinely contradict old 
truth. Tills may not at first be clear. New ranges 
of experience such as those due to the invention 
of the telescope and microscope have actually led 
to alteration of our categories in historic time. 
The same thing may happen through more pene- 
trating or adequate analysis of old types of ex- 
perience witness Virchow's redefinition of dis- 
ease. What was previously regarded as real e. g. 9 
disease entities may come to be looked upon as 
unreal, and what was previously taken to be un- 
real e. g., curved space may be admitted to 
reality. But when this happens the truth remains 
unaltered and new truth and old truth do not 
contradict. Categories and concepts do not liter- 
ally change; they are simply given, up and re- 
placed by new ones. When disease entities give 
place to mere adjectival states of the organism 
induced by changed conditions such as bacteria, 
the old description of the phenomena of disease 
does not become false in any sense in which it was 
not always false* All objects are abstractions of 
one sort or another; a disease entity is found to 
be a relatively poor kind of abstraction for the 
understanding and control of the phenomena in 
question. But in terms of this abstraction any in- 
terpretation of experience which ever was cor- 
rectly made will still remain true. Any contradic- 
tion between the old truth and the new is 'verbal 
only, because the old word "disease" has a new 


meaning. The old word is retained but the old 
concept is discarded as a poor intellectual instru- 
ment and replaced by a better one. Categories 
and precise concepts are logical structures, Pla- 
tonic Ideas; the implications of them are eternal 
and the empirical truth about anything given, ex- 
pressed in terms of them, is likewise through all 
time unalterable. 

In the typical case in which old methods of in- 
terpretation are discarded in favor of new ones, 
it requires new empirical data, which offer some 
difficulty of interpretation in the old terms, to 
bring about the change- Any set of basic concepts 
has vested interests in the whole body of truth 
expressed in terms of them, and the social prac- 
tices based upon them. The advantage of the 
change must be considerable and fairly clear in 
order to overcome human inertia and the prestige 
of old habits of thought. Such new and recalci- 
trant data, which bring about the change, com- 
plicate the problem of comparing the "new truth" 
with the old. The factors which need to be con- 
sidered are: (1) the two sets of concepts, old and 
new, (2) the expanding bounds of experience in 
which what is novel has come to light, (3) the 
conditions of the application of the concepts to 
this new body of total relevant experience. 

In the case of the Copernican revolution, for 
example, it was the invention of the telescope and 
the increasing accuracy of observation which 


mainly provided the impetus to reinterpretation. 
But these new data were decisive only in the prag- 
matic sense. Those who argued the issue supposed 
that they were discussing a question of empirical 
fact. But since there is no absolute motion, the 
question what moves and what is motionless in the 
heavens is one which cannot be settled by experi- 
ence alone- The fixed stars prove to be a highly 
convenient frame of refer ence 5 resulting in rela- 
tively simple generalizations for the celestial mo- 
tions, and enabling celestial and sublunary phe- 
nomena to be reduced to the same equations, while 
almost insurmountable complexity and difficulty 
attend the choice of axes through the earth. Theo- 
retically, however, if any system of motions is 
describable with respect to one set of axes, it is 
also describable in terms of any other set which 
moves with reference to the first according to any 
general rule. Let us imagine for a moment that 
this theoretically possible description of astro- 
nomical and physical phenomena in terms of a 
motionless earth had been worked out for all the 
data now at hand. In terms of which set of con- 
cepts, old or new, should we have the truth? Ob- 
viously in both. The one would be comprehension 
and simple truth; the other so complex as to be 
almost or quite unworkable. But they would no 
more contradict each other than a measurement 
in pounds and feet contradicts one in grams and 


This situation is not altered by any thought 
that newly discovered fact may play another than 
the pragmatic role and be decisive of truth in a 
deeper sense. Nobody has ever supposed that what 
were only hypotheses or empirical generalization 
of a high degree of probability were incapable of 
being disproven by new facts. To the extent that 
newly discovered empirical evidence may render 
old principles theoretically impossible, the old 
truth never was anything but hypothesis and is 
now proved flatly false. It is not, I hope, the point 
of the pragmatic theory of knowledge to reduce 
all truth thus to hypothesis. That would be noth- 
ing but a cheerful form of skepticism. 

Rather the point of the pragmatic theory is, I 
take it, the responsiveness of truth to human bent 
or need, and the fact that in some sense it is made 
by mind. From the point of view here presented, 
this is valid, because the interpretation of experi- 
ence must always be in terms of categories and 
concepts which the mind itself determines. There 
may be alternative conceptual systems, giving rise 
to alternative descriptions of experience, which 
are equally objective and equally valid, if there 
be not some purely logical defect in these cate- 
gorial conceptions. When this is so, choice will 
be determined, consciously or unconsciously, on 
pragmatic grounds. New facts may cause a shift- 
ing of such grounds. When historically such 
change of interpretation takes place we shall gen- 


uinely have new truth, whose newness represents 
the creative power of human thought and the 
ruling consideration of human purpose. 

The separation of the factors, however, reveals 
the fact that the pragmatic element in knowledge 
concerns the choice in application of conceptual 
modes of interpretation. On the one side, we have 
the abstract concepts themselves, with their purely 
logical implications. The truth about these is ab- 
solute, in the fashion in which pure mathematics 
offers the typical illustration. Such purely ab- 
stract a priori truth answers only to the criteria 
of consistency and adequacy. It is absolute and 
eternal. On the other side, there is the absolute 
brute-fact of given experience. Though in one 
sense ineffable, yet the given is its own fashion 
determinate ; once the categorial system, in terms 
of which it is to be interpreted, is fixed, and con- 
cepts have been assigned a denotation in terms 
of sensation and imagery, it is this given experi- 
ence which determines the truths of nature. It 
is between these two, in the choice of conceptual 
system for application and in the assigning of 
sensuous denotation to the abstract concept, that 
there is a pragmatic element in truth and knowl- 
edge. In this middle ground of trial and error, of 
expanding experience and the continual shift and 
modification of conception in our effort to cope 
with it, the drama of human interpretation and 
the control of nature is forever being played. That 


the issues here are pragmatic; that they do not 
touch that truth which still is absolute and eter- 
nal this is the only thing that would save those 
who appreciate the continually changing char- 
acter of this spectacle from skepticism. 


So far, various aspects of, and elements in, 
knowledge have been considered, for the most part 
separately. The results may now be brought to- 
gether, and the relation of the various factors 
considered with more care. Quite commonly, the 
different types or phases of knowledge of pre- 
sented objects, of a priori principles, of empirical 
generalizations are all lumped together, and 
either what is true of one type only is applied to 
all, or in some other fashion an omnibus explana- 
tion is attempted which will include them all under 
one formula. No theory devised by such procedure 
can ever be more than partially successful. 

The following need to be distinguished: (1) the 
immediate awareness of the given, such as might 
be supposed to be reported in statements like 
"This looks round," "This feels hard," "This 
tastes sweet"; (2) knowledge of presented ob- 
jects, such as is expressed by, "This is hard," 
"This is a sweet apple," "This penny is round" ; 
(3) the a priori elaboration of wholly abstract 
concepts, like the formulations of pure mathe- 
matics, apart from any question of possible ap- 
plications; (4) the categorial knowledge of inter- 


pretative principles and criteria of reality, which 
is that form of a priori knowledge which arises 
when concepts have a fixed denotation and are ap- 
plied to the given; (5) empirical generalizations, 
which are universal but not a priori. 

The difference of the first two is especially easy 
to overlook; the first "This looks round" is a 
direct report of the momentarily given; the sec- 
ond "This penny is round" represents the con- 
ceptual interpretation of this given, and implies 
much which is not given. 

As has been made clear in preceding chapters, 
the term "knowledge" is here restricted in a man- 
ner which may seem arbitrary but at least answers 
to definite criteria; that only is called "knowl- 
edge" which is verifiable and has a significant op- 
posite "error," There is, in the knowledge of pre- 
sented objects and of objective properties, a 
distinguishable element of awareness which is in- 
dispensable to this knowledge but which, by itself, 
cannot be knowledge in the meaning here as- 
signed. It is this awareness of immediate and rec- 
ognizable qualia which may be supposed to be 
expressed by "This looks round," etc. So far as 
such a statement is merely a report of the immedi- 
ately given, it is neither verifiable nor stands in 
any need of verification. Nor is it subject to any 
possible error. If the subject of the report is the 
immediate feeling itself, there is no possibility of 
being mistaken about it, though all sorts of mis- 


apprehension about the experience reported might 
arise in the mind of another person to whom the 
report might be addressed, due to poor choice of 
language or to the ineffable character of individ- 
ual immediate experience. 

It has also been pointed out that such immedi- 
ate awareness is an element in knowledge rather 
than a state of mind occurring by itself or pre- 
ceding conceptual interpretation. Without the re- 
lational element which conception introduces, im- 
mediacy is inarticulate. It is questionable whether 
a state of pure esthesis is a genuine mental pos- 
sibility. In any case, it is not necessary to assume 
its existence. The sense in which such immediacy 
is prior to its interpretation is the sense in which 
interpretation is subject to change. Cognition of 
the external world is active just so far as con- 
ceptual interpretation is subject to correction and 
alteration, on occasion. In the case of such a new 
interpretation, the immediate awareness is liter- 
ally and temporally antecedent ; but that there is 
a first moment of such apprehension, in which 
there is such awareness and no interpretation, it 
is not necessary to believe. In all cases, however, 
it is the content of the given which determines (in 
part) the interpretation, not the interpretation 
which determines the immediate to fit it. In that 
sense also the awareness of the given is prior to 
its interpretation. 

Predications of the second sort "This is hard," 


"This penny is round" express something much 
more complex than the immediate feeling of pres- 
sure or the givenness of a "round" or "elliptical" 
presentation. As predications of objective prop- 
erties, these represent an interpretation put upon 
the content of immediate awareness which im- 
plicitly predicts further experience. Being thus 
predictive, they are judgments which are subject 
to verification and liable to error. 

Strictly, the nature of the judgment as inter- 
pretation of the given is more complex than the 
statement of it. This is obvious in the case of the 
example, "This penny is round**; calling what is 
presented "a penny* 5 is interpretive to at least the 
same degree as judging it to be round. What the 
subject expresses is an interpretation as much as 
the predicate. A similar thing is true of all predi- 
cations in which the subject is something named; 
the distinction between subject and predicate is 
merely one of emphasis or nuance it reflects a 
difference between what is already granted or ob- 
vious and what needs explicit assertion. The two 
statements, "This penny is round" and "This is 
a round penny/* do not represent materially dif- 
ferent interpretations of what is given. 

Similar considerations hold for all predications 
of objective character to anything given, except 
those in which the subject is mere demonstrative 
like "this" which supposedly serves merely to 
point to the given. Even such demonstratives, 


moreover, attribute to the "this" which is indi- 
cated that kind of continuance and lawful rela- 
tion to other things which is the mark of the ob- 
jective* The point of using the demonstrative lies 
in its indication of a reality which is not subjec- 
tive but whose existence and asserted character is 
verifiable, by ourselves and other persons, over a 
stretch of time. Hence the difference between as- 
sertions about a "this" and predications about 
something named lies only in the more extended 
and more definite character of the interpretation 
which the use of a name as the subject term im- 
plicitly takes for granted. And in all such predi- 
cation the application of the subject term, as of 
the predicate, is an interpretation of given ex- 

The fact that the denotation of a demonstrative 
is, in all ordinary parlance, "this (thing or ob- 
jective property) which is presented," and not 
"this presentation^ draws our attention to the 
fact that the attempt to exemplify the mere re- 
port of immediate awareness as by the illustra- 
tions, "This tastes sweet," etc. can be only ap- 
proximately successful. Here also the demonstra- 
tive denotes an object, not the presentation itself. 
Otherwise there would be no predication, because 
subject and predicate would coincide and what we 
are trying to express would take the form "Sweet 
taste tastes sweet" or simply the ejaculation 
"Sweet taste!" The use of predication is com- 


pletely pre-empted to the conveying of the objec- 
tive, and there is no language whatever, unless of 
primitive cries, which expresses awareness of the 
given as such. The expressive ejaculation or cry 
is, so to speak, the salutation of the real or the ac- 
knowledgment of an existent. Such salutation 
or acknowledgment, however, represents an essen- 
tial element in knowledge that which distin- 
guishes the truth from lies ; that fact of presenta- 
tion which is our confrontation with reality. 

Concerning our knowledge of presented objects, 
it may be noted further that there is no difference 
which is important in the present connection be- 
tween the predication of an objective property 
and the predication of substantive character or 
thinghood. So much is evident from the practical 
equivalence of "This penny is round" and "This 
is a round penny." The kind of interpretation in- 
volved and the general manner of its verification 
is the same whether it is a property like round- 
ness or a substantive character such as being a 
penny which is in question. 

By how much does the interpretation which 
characterizes our knowledge of objects transcend 
what is given? What is involved in its complete 
verification? Obviously in the statement "This 
penny is round" I assert implicitly everytlwng the 
failure of which would 'falsify the statement. The 
implicit prediction of aU experience which is es- 
sential to its truth must be contained in the origi- 


nal judgment. Otherwise such experience would 
be irrelevant to it. All that further experience the 
failure of which would lead to the repudiation of 
the apprehension as illusory or mistaken is pre- 
dicted in the judgment made. Now suppose we ask : 
How long will it be possible to verify in some man- 
ner the fact that this penny is round? What total- 
ity of experience would verify it completely, 
beyond the possibility of necessary reconsidera- 
tion? I have here no theoretical axe to grind, but 
it seems to be the fact that no verification would 
be absolutely complete ; that all verification is par- 
tial and a matter of degree. Even after the penny 
itself has ceased to exist, a sufficiently important 
connection with other matters might still lead to 
a revival of the question whether it is really true 
that a round penny lay before me on my desk on 
the twenty-ninth of January, 1926, at four 
o'clock And however difficult it might be at such 
later time to gain new and decisive evidence, theo- 
retical tests of what would increase or decrease 
the probability would be capable of formulation, 
Is it not the case that the simplest statement of 
objective particular fact implicitly asserts some- 
thing about possible experience throughout all 
future time; that theoretically every objective 
fact is capable of some (partial) verification at 
any later date, and that no totality of such ex- 
perience is absolutely and completely sufficient to 
put our knowledge of such particulars beyond all 


possibility of turning out to be in error? So far 
as this is true, all interpretation of particulars 
and all knowledge of objects is probable only, 
however high the degree of its probability. Every 
such judgment about the real external world re- 
mains forever at the mercy of future possible ex- 
perience. Between the immediate awareness, "This 
looks round, 5 ' and the objective interpretation, 
"This is round," there lies all the difference be- 
tween this present moment and all time; between 
an experience which is now complete and Tiad, and 
a totality of possible experience which is unlim- 
ited and inexhaustible. 

Perhaps this conclusion that the verification of 
empirical particulars is always partial, will be 
more evident if we investigate the alternative. 
Let us suppose that, for some interpretation of a 
particular given experience, there is some finite 
and limited totality of later experience which, to- 
gether with the presentation itself, will absolutely 
and finally verify our interpretation as an em- 
pirical fact so that nothing further could possibly 
require the retraction of it. Let us suppose the 
presentation to be given and interpreted at time 
*Q. And let us imagine the verification to be com- 
plete at time t. Now let us ask: Will the total 
significance of this empirical fact be exhausted 
at time t 9 so that consideration of it will enable 
no further prediction of anything which would be 
discoverable at some still later time, t%? We shall 


hardly be inclined to answer in the affirmative, 
since such an answer would mean that, after the 
date ti 9 the fact would have no further conse- 
quences and its existence or historical verity 
would be no longer determinable. Because if at 
time t% there should be no consequences of this 
fact, which would be lacking if the event had not 
taken place, then there will be absolutely no 
means, at time t^ whereby it may be discovered 
whether there was any such fact or not. Hence 
there can be no such time, t 9 after which no con- 
sequences predictable from the interpretation of 
the presentation at *o will accrue. The verifiable 
consequences of any fact last as long as time 
itself. Now suppose further that at some date, t^ 
we put ourselves in position to meet the conse- 
quences of this fact, which was accepted as com- 
pletely established at t. And suppose that these 
consequences fail to appear, or are not what the 
nature of the accepted fact requires? In that case, 
will there still be no doubt about the accepted 
fact? Or will what was supposedly established at 
i be subject to doubt at 2 ? And in the latter 
case can we suppose it was absolutely verified at 
time #1? Since no single experience can be abso- 
lutely guaranteed to be veridical, no limited col- 
lection or succession of experiences can absolutely 
guarantee an empirical fact as certain beyond the 
possibility of reconsideration. 

This is not an argument; as argument, it would 


be petitio principiL But I believe that, having 
followed out these consequences, the reader will 
be inclined to repudiate the supposed alternative 
and to grant the thesis that the interpretation 
put upon the empirically given is always some- 
thing temporally Inexhaustible, hence never com- 
pletely verified, and is always probable only. 

Incidentally this conclusion serves to answer a 
question which may have occurred to the reader 
in an earlier chapter: If knowledge is knowledge 
only as it is predictive and verifiable, does it cease 
to be knowledge when it is completely verified? 
We now see the answer to be that knowledge the 
knowledge of empirical particulars at least never 
is completely verified. This also throws light upon 
the nature of the concepts of things and of ob- 
jective properties, which are applied to the given 
in interpreting it. It has been noted that what 
any concept denotes has always a temporal spread. 
In appealing to the reader to grant the consid- 
erations of the preceding paragraphs, I likewise 
ask him to grant, by implication, that this tem- 
poral spread is unlimited and that the existence 
of things and properties has a sort of unlimited 
duration, as was suggested in Chapter IV in the 
reference to the "fallacy of simple location" in 

The empirical knowledge of particular objects 
is probable only, Yet underlying it there is and 
must be an element of the a priori, else there 


would be no criteria of its empirical truth and not 
even a valid probability. When we make the judg- 
ment, "This penny is round, 55 the subject "penny" 
and the predicate "round" both express implicitly 
certain a priori criteria which are definitive of 
the meaning of these terms in application. Being 
a penny, or being round, means a hundred and 
one various sequences in further possible experi- 
ence. That these sequences would actually accrue 
under suitable conditions is implicitly predicted 
in applying these concepts to the given presen- 
tation. To be really round, this presented object 
must alter in appearance in certain characteristic 
ways with change of perspective; it must feel in 
certain ways if handled; and if it be measured 
with precision instruments, the results must be 
thus and so. Furthermore, it must not alter in cer- 
tain other ways which would oblige us to explain 
that the angle of vision or some trick of eyesight, 
etc., had led us into error. Such explication of 
what is implicit in the concept sets the criteria by 
which further experience will verify (or falsify) 
the present judgment. If this setting of criteria 
were not a priori and incapable of being over- 
turned by the eventualities of experience, then 
such experience would not be a test of the truth 
of the predication, or even establish a probability 
of it. That is to say, when we make the judgment, 
"This is round," what we suppose ourselves to 
know requires two propositions to express it fully : 


(1) "If this is round, then further experience of 
it will be thus and so (the empirical criteria of 
objective roundness)" and (2) "This present 
given is such that further experience (probably) 
will be thus and so." The first of these is a priori ; 
the second is our statement of the probable em- 
pirical truth about the given object. 

Unfortunately, I must ask the reader to go 
with me a little further in this analysis of the 
knowledge of empirical particulars. There is no 
escape for us, because the truth about it happens 
to be quite complex. Certain further considera- 
tions have, in all probability, already suggested 
themselves to the reader. First, in the light of pre- 
ceding paragraphs it is evident that the empirical 
criteria of, let us say, roundness which we can 
have explicitly in mind at any one moment, will 
not be the completely sufficient criteria because 
the empirical eventualities which would completely 
verify the roundness of the given thing are too 
utterly complex and extended to be thus explicit* 
When we actually define "roundness" in words, 
we refer the meaning of one term to others, and 
(supposing the definition to be well-made and ade- 
quate) both the subject term defined and the 
predicate which defines it have a meaning which, 
in application, are of this complex sort which it 
is not possible to express completely, or have ex- 
plicitly in mind, at any one moment. So to speak, 
the terms of a definition are all in one plane, in 


which the definitive relations are finitely specifi- 
able. But each term is a point of reference in an- 
other dimension (that of denotation) and its 
meaning in this dimension is not thus exhaustively 
expressible in a few words. That a meaning in 
denotation must have this complex character is 
due in part to the fact that no object can be ap- 
prehended in isolation. Any appearance of an ob- 
ject is conditioned also by other objects, particu- 
larly by my own body. It is such conditions which 
are expressed in the "if" clause of those "If 
then 55 propositions in which the predictions im- 
plicit in an interpretation may be made explicit. 

That such complexity of consequences is always 
implicit in every concept, may be more evident if 
we reflect upon the example of some mathematical 
concept. For instance, the analytic consequences 
of the definition of "triangle" are as numerous as 
the geometrical theorems concerning triangles 
that is, unlimited in number. 

We may further note that we may indeed 
must have this unlimited meaning in denotation 
in mind in a figurative sense, when we do not have 
it in mind explicitly. The possession of the con- 
cept or meaning is utterly irrelevant to any ex- 
perience whatever, unless we could with sufficient 
time and attention at least say whether certain 
eventualities of experience would or would not be 
compatible with its application. For example, the 
detective who should be assigned the problem 


whether I did or did not have a round penny be- 
fore me at a certain time, could not possibly ex- 
haust in any statement all the evidence which 
would, if discovered, be pertinent to it. But if he 
did not have it in mind in the figurative sense, all 
the Sherlock Holmesing in the world would not 
help him a bit because he would not be able to 
recognize evidence when he found it. We are, even 
the best of us, more or less stupid and might not 
be capable of specifying whether a given eventu- 
ality is or is not pertinent to our problem. That 
is to say, it may be true in some degree that we do 
not have the meaning of our concepts in mind 
even in the figurative sense* But so far as we are 
thus stupid, we simply do not know what we mean 
when we use the terms; that part of the (ideal) 
meaning of the language we apply to our pre- 
sented experience is simply erased in our case; 
and we are condemned by our lack of intelligence 
not to understand what we are trying to think and 
talk about. 

However, this stupidity which in some degree is 
inevitable to all of us, and this fact that much of 
our meaning., when we interpret given experience, 
is in mind only in the figurative sense, is not par- 
ticularly embarrassing, either theoretically or 
practically. Because while the empirical eventual- 
ities which would exhaust the denotative meaning 
of our interpretative concept are too numerous to 
state or have in mind explicitly, those which we 


shall be called upon to meet in any brief period 
are not. And we can with wit and luck keep 
ahead of experience and be explicitly prepared to 
interpret it by the time the evidence is presented. 
Furthermore, such explicit eventualities as we do 
have in mind represent knowledge which is a pri- 
ori, however far they fall short of exhausting our 
interpretive meaning, 

A second point which needs at least passing 
mention is the fact that the a priori element in 
our knowledge of particulars is complex in an- 
other fashion also. This element has been ex- 
pressed above by the proposition "If this is round, 
then , . ." So far we have observed that the 
"then" clause is compounded of empirical eventu- 
alities apparently unlimited in number. We must 
now observe, further, that each one of these is 
conditional. "If this is round, then if I take two 
steps to the right, it will look more elliptical." "If 
this is round, then if it be measured with precision 
instruments, the result will be thus and so. 55 "If 
this is round, then it will not look elliptical if it is 
viewed from directly in front* This conditional 
character of the propositions expressing the pre- 
dictions implicit in the interpretation of the given, 
has already been discussed in Chapter IV; and 
hence need not detain us longer here. 

These first two points may be schematically 
summed up by saying that the a priori element 
which must underlie the empirical knowledge of 


particulars is expressible by some set of proposi- 
tions of the form: If this is round, then condition 
A being provided, empirical eventuality M will 
accrue. If this is round, then condition B being 
provided, empirical eventuality N will accrue. And 
so on. The totality of the complex "then" classes 
expresses the complete and a priori meaning of 
the concept "round" in denotation. Some of these 
we have explicitly in mind granted intelligence, 
we shall be prepared with those with respect to 
which experience will shortly present us pertinent 
evidence. Others we have in mind in the figurative 
sense that we should be explicitly prepared if 
called upon to include them in, or exclude them 
from, what we mean by "being round." Still 
others, perhaps, which belong to an ideal denota- 
tion of "roundness" we may not have in mind at 
all. So far as this last is true, we are stupid and 
are unprepared to understand experience. 

There is still one most important point about 
our knowledge of empirical particulars which re- 
mains to be observed. It has been said that what 
we must know, for example, in order to interpret 
a given presentation as that of a round object is 
(1) the a priori proposition, "If this is round, 
then further experience of it will be thus and so, 5 ' 
and (2) "This present given is such that further 
experience (probably) mil be thus and so. 5 * It will 
have occurred to the reader that we could not pos- 
sibly know anything like the second proposition 


except by a generalization from previous experi- 
ence "Things which look as this does, under con- 
ditions like the present, usually turn out to satisfy 
the criteria of roundness in further experience." 
In general, we must rely upon past experience for 
our knowledge of what empirical eventualities are 
likely to be connected with any given type of ap- 
pearance that is, for our knowledge as to what 
presentations are appearances of what objects. In 
every case, interpretation of the given requires 
empirical generalization. 

But it is particularly necessary to recognize 
that the subject of this type of generalization 
from experience is not the object presented (or 
the class of such objects) but the presentation it- 
self (or the class of such). Our recognition of the 
object "This is a round penny" is the inter- 
pretation itself. But the recognition of the pres- 
entation is simply the classification of it with other 
qualitatively similar appearances. The basis of 
our interpretive judgment is the fact that, in past 
experience, what appeared as this does, under cir- 
cumstances like the present, has turned out to be, 
for example, a round penny in a sufficiently 
large proportion of cases to warrant probable 

This kind of generalization is seldom explicit. 
Our collation of the given with similar appear- 
ances in the past is too swift and instinctive for 
that. This is presumably the element in human 


knowledge which is evolutionally basic and is 
shared bv us with the other animals. But occasion- 


ally at least it needs to be explicit; and there is 
one familiar example which may serve as illustra- 
tion. The physician is often called upon to diag- 
nose a case, not by some decisive test, but by "the 
picture*' which the case presents. (A clinical pic- 
ture is, of course, much more complex than any 
single presentation, but it is, so to speak, preemi- 
nently presentational in character, requiring to 
be identified by direct inspection and difficult or 
impossible to put in words.) If the physician de- 
cides, "This is measles," he does so by means of a 
generalization from past experience. The subject 
of this generalization is the class of clinical pic- 
tures like the present one. The basis of the judg- 
ment is the frequency with which such appearances 
in the past have been followed by a later case- 
development which answers to the elaborated med- 
ical concept "measles." The ascription of this 
concept is an interpretation, whose main signifi- 
cance is that of prognosis. 

Our knowledge of objects in general is such a 
diagnosis of appearances. Without our recogni- 
tion of the presentation by its classification with 
qualitatively similar ones in the past, and our 
recollection of the further eventualities in these 
cases, no interpretation would be possible. That 
such recognition is ordinarily spontaneous and 
unconsidered, does not alter the logical character 
of it as a generalization. 


In empirical knowledge there are, thus, two ele- 
ments concerning which we have certainty; the 
recognized qualitative character of the given pres- 
entation is one and the a priori elaboration of 
some concept (or the partial elaboration which is 
explicit with us) is the other. But the applicabil- 
ity of the concept to the presentation is probable 
only, because such application is an interpreta- 
tion which is predictive. The probability, or degree 
of assurance of such interpretation, reflects a gen- 
eralization from experience. It is an argument 
from past to future or from the uniformity of 
experience. In short, in spite of the elements of 
certainty underlying it, which have been pointed 
out, the validity of our knowledge of presented 
objects depends upon the same general considera- 
tions that govern empirical generalizations of the 
types more frequently recognized. Oftentimes it 
has been supposed that we have a kind of cer- 
tainty in our recognition of objects, and this has 
been designated as "knowledge by acquaintance." 
But such is not the case. What we are directly 
"acquainted with" are not objects but presenta- 
tions. Seeing is believing, but quite often the be- 
lief based on seeing turns out to be a false one. 

Thus there are, on the theory here presented, 
very much the same problems to be met concern- 
ing empirical knowledge, as on various other the- 
ories* But the incidence of those problems is dif- 
ferent; and this, as we shall discover in later 


chapters, makes a difference in the manner in 
which they can be met. A brief comparison, even 
though inadequate, may be of assistance here. 
Ordinarily the problem of empirical knowledge in 
general, when phrased by reference to usual forms 
of predication, is viewed as the problem how we 
can know with certainty that what is denoted by 
the subject of the proposition whose truth is in 
question will also have the properties denoted by 
the predicate. How may we be assured that all 
JST's are F's; that all masses obey Newton's law; 
that all events have causes, that the sum of the 
angles of a triangle will be 180 degrees? The prob- 
lem is that of the necessary connection of what is 
denoted by two concepts. 

Now I believe it should be obvious, no matter 
what theory of knowledge is held, that there is 
only one ground on which the necessary connec- 
tion of X and F, such that all X*s will certainly 
be F ? s can be known ; that is, that if we find that 
the concept F is inapplicable to any particular, 
then the concept X will be retracted as likewise 
inapplicable. If we know with certainty in advance 
that all men are mortal, we know it because if we 
discover any being not to be subject to the acci- 
dent of death, then, however like a man. he may 
appear, we shall refuse to recognize him as hu- 
man. We can know that the sum of the angles of 
a triangle must be 180 degrees only if in case we 
find this not so in a particular instance, we shall 


retract the concept "Euclidean triangle 55 as in- 
applicable to the thing in question. Whenever this 
is so, the subject concept implies or includes the 
predicate concept and the proposition is a priori 
because the judgment is analytic. 

If the problem of empirical knowledge is sup- 
posed to be "How can we be assured that objects 
met with in experience and identified as JST's will 
never turn out not to be F*s?" then, again, I be- 
lieve it should be clear on any theory of knowledge 
that we can never have such, assurance, especially 
if it is remembered that there is no a priori cer- 
tainty that perception in any particular case is 
not illusory or subject to the accident of mistaken 
identification* That is to say, it should always have 
been clear, if we remember that "experience" as 
we are confronted with it includes the non-veridi- 
cal, that the application of concepts in naming 
and recognizing objects, itself implies characters 
of the object which are not now presented but 
wait upon further experience to be revealed. 

Remembering this, it will be evident that the 
only kind of a priori knowledge of the empirical 
for which there is room in a consistent theory is 
that kind which consists in knowing the empirical 
eventualities, implicit in the application of our 
subject-concept, which are indispensable to the 
correctness of such application. We do not need 
any limitation of possible experience to be assured 
of this. If a thing is not an X unless further ex- 


perience will corroborate a certain F-character of 
it, then, let experience be what it will, all JST's must 
be F's. It is true that this leaves all empirical 
knowledge except the hypothetical subject to 
the tests of later experience, in the sense that such 
future experience may invalidate our identifica- 
tion and naming of objects. But my point is that 
no theory ever presented can do more, unless we 
can be absolutely certain of our recognitions of 
objects in momentary experience. Such certainty 
of momentary identification is something which no 
theory has ever claimed explicitly, though most 
theories have, as a matter of fact, proceeded as if 
it were possible. 

Take, for example, geometrical knowledge. 
Kant claimed a synthetic ground of our knowl- 
edge of the necessary connection between "the 
triangular" and "figures having interior angles 
which sum up to 180 degrees." This synthetic 
ground was supplied by the limitation of possible 
human experience to the Euclidean space-form. 
But did Kant suppose that this limitation of pos- 
sible experience enabled us to glance at a plot of 
ground and say with a priori certainty that the 
sum of its angles was exactly 180 degrees? If not, 
could one claim accepting Kant's theory that 
this a priori limitation of experience in any fash- 
ion assured to our knowledge of this particular 
plot of ground (or of any other particular) a 
certainty which future experience might not 


trench upon? There can be no doubt that the an- 
swer must be negative. What, then, as regards our 
knowledge of the empirical, is the supposed ad- 
vantage of this synthetic element supplied by the 
limitation of possible experience? Absolutely all 
that it assures us, about any empirically given 
object, is that, if it be truly triangular, then the 
sum of its angles will be 180 degrees ; and whether 
this condition is satisfied in any particular case, 
it leaves subject to the test of future and more 
exact determination. 

One may further inquire: If we can never, in 
a momentary apprehension of anything, be abso- 
lutely certain that it is a real Euclidean triangle, 
how are we to know that the class of objects actu- 
ally identifiable as triangles have this further 
property of the sum of their angles, unless by the 
fact that if they do not have it, then we shall re- 
tract the ascription of real triangularity as in- 
applicable to them? We now see this plot of 
ground: it looks triangular. In this momentary 
experience, there is absolutely no way in which we 
can tell that it is not. Suppose we proceed to mea- 
sure its angles with a theodolite, and find that the 
sum is 181 degrees. Does the Kantian, or any 
other, theory deny this possibility? If not, then 
the only ground on which we can know that what 
is actually identifiable as triangle will have the 
further property in question is, in the last anal- 
ysis, some ground which will lead to the repudia- 


tion as **not-triangular" of that which fails, in 
further experience, to give the required measure 
of its interior angles^ even though, as presented 
at a certain moment, it was indistinguishable from 
a real triangle. 

Let us take one step further. If a good Kantian 
must still grant, in the interests of veracity, that 
the generalization "All figures which are momen- 
tarily indistinguishable from the triangular have 
angles which sum up to 180 degrees" would be 
false, then how can it be claimed with plausibility 
that experience of a world in which the laws of 
Euclid do not hold would be impossible to human 
beings? If in one case we can see as triangular a 
figure the measure of whose angles exceeds 180 
degrees, in what sense would it be impossible that 
this should be universally true? How would a 
form of intuition prevent us from seeing just as 
we now do a space whose properties would be Rie- 

Is it not fairly obvious that the real question 
about the applicability of Euclid to our space 
is not the question of what we can imagine or 
could conceivably experience; but is the question 
whether the character of experience in general is 
such that the procedure by which the failure of 
any spatial object to verify the Euclidean prop- 
erties is put down to mistaken apprehension is one 
which gets on better (L e., better serves our inter- 
ests of reducing experience to order and securing 


control) than would any general revision of our 
whole system of geometrical conceptions? 

The situation in which we find ourselves as a 
result of modern developments in geometry is in 
accord with this conception of a priori knowledge 
and its relation to the empirical. As modern 
mathematics discovers, the concept "triangle" 
(and the other basic concepts of geometry) in- 
clude a logically sufficient ground for all the 
properties of triangles, without any synthetic ele- 
ment which is supposed to limit experience. To 
be sure, this means entire separation of the ques- 
tion of abstract geometrical truth from the truth 
of experience. This last becomes a question of 
empirical generalization or more accurately, it 
is the complex question: Which of the alternative 
systems of geometrical concepts will best succeed 
in its application to experience; it being remem- 
bered that what such application requires is that 
any geometrical concept will be retracted as in- 
applicable when experience fails to verify the es- 
sential geometrical properties L e.> that all di- 
vergence of experience from our chosen geometry 
is to be explained on other grounds, or relegated 
to the status of mistaken apprehension? 

From this point of view, the development of 
the conceptual system in the abstract is a priori ; 
the question of the applicability of one of its con- 
stituent concepts to any single particular is a 
matter of probability ; and the question of appli- 


cation in general is the question of the choice of 
an abstract conceptual system, determined by 
pragmatic considerations. 

We find that there is much more in the ab- 
stract concepts than Kant thought that as a 
fact the whole geometrical system can be drawn 
from them by purely logical analysis ; and we find 
that in any case the application to particulars is 
no better than probable. Under these circum- 
stances, it is extremely dubious what advantage 
would accrue if we could find a ground for a priori 
truth which was synthetic and consisted in some 
limitation of the possibilities of intuition. And we 
find the supposition that there is limitation and 
that we can know it, vitiated by the fact that we 
most certainly could have an experience in which 
Euclidean-appearing 1 things should, upon fur- 
ther examination, turn out to have non-Euclidean 
properties. The only question about a priori truth 
in application which is left to be determined is 
the question what shall be accepted as the em- 
pirical criteria of triangularity, straightness, and 
so on. This is at once the question what kinds of 
sequences in experience are to be regarded as 
ground for attributing mistake to previous identi- 
fications of spatial characters in things, and the 
question what abstract system shall be our choice 
for application to experience in general. The 
chosen system becomes criterion of the veridical 
in experience, that is, its concepts become criteria 


of reality of a certain, sort- It is this question of 
the choice of conceptual systems for the inter- 
pretation of experience which, on the view here 
presented, is a matter of pragmatic choice, 
whether that choice be made deliberately, or un- 
consciously and without recognition of its real 

If, now, the reader will generalize from this 
illustration in terms of the geometrical, he will 
have before him the distinguishing characteristics 
of the present theory of the relation between a 
priori truth and the content of experience. While 
other concepts than the mathematical do not usu- 
ally have their consequences so systematically 
worked out, nevertheless all concepts give rise to 
an a priori truth which is purely analytic and in- 
dependent of any application to experience. Such 
analytic consequences of a single concept, in iso- 
lation, will be relatively meager and relatively 
trivial. But how complex, far-reaching, and im- 
portant the analytic consequences may be when 
three or four such abstract concepts are con- 
joined, modern systems of mathematics serve to 
illustrate. That which any such concept denotes 
is always something which, in terms of experience, 
must have a temporal spread. What is required to 
determine its applicability, is some orderly se- 
quence in experience, or some set of such. At any 
given moment, such applicability is verifiable 
only approximately or in degree. It is thus that 


the application of the concept to experience may 
be secured without loss of its a priori character: 
its logical consequences, which time alone can 
verify, become criteria of its applicability. Later 
experience which does not accord will lead to the 
retraction of the concept as inapplicable to the 
particular to which it was assigned by previous 
interpretation. Thus the logical requisites a pri- 
ori of the concept become, in its application to 
experience, the criteria of reality of a certain sort. 
The application a priori of Euclidean geometry 
to nature means, for example, that whatever ap- 
prehended particular turns out, in the course of 
experience, not to have the properties logically 
implied by the concept "Euclidean triangle" will 
be condemned as not a real triangle^ however much 
it may have looked like one.* 

Up to this point, there has been no considera- 
tion of the last of the five distinct kinds of ap- 
prehension or knowledge mentioned at the begin- 
ning of the chapter, L e. 9 empirical generaliza- 
tions, ordinarily so-called. We have found that a 
certain kind of empirical generalization enters 
into the judgment of truth about empirical par- 
ticulars. But these are not of the type ordinarily 
called "generalizations" since the subject of them 
is the presentation itself; they are usually not 
expressed at all and are indeed, as we found, dif- 

*For a note concerning a further problem about the application of 
abstract conceptual systems to experience, see Appendix E. 


ficult to express in language without including 
reference beyond themselves to objective and en- 
during things. Customarily what is meant by "em- 
pirical generalization" is a universal proposition 
the subject of which denotes a class of objects. It 
is distinguished from the a priori in application 
by the fact that the connection between subject 
and predicate is not necessary but contingent.* 
A simple illustration may be here of service. The 
proposition "All swans are birds" is a priori be- 
cause if any creature originally designated as a 
"swan" should be discovered to lack some dis- 
tinguishing character of birds, the name "swan" 
would be withdrawn. The applicability of the 
predicate term is logically requisite to the ap- 
plicability of the subject. But the proposition 
"All swans are white" is an empirical generaliza- 
tion because white color is not included as essen- 
tial in the denotation assigned to "swan." The 
former proposition can not be falsified by any 
possible experience because its truth has a purely 
logical warrant; it represents the implication of 
a concept. But the latter proposition has no such 
logical warrant and may be falsified by experi- 
ence ; black creatures having all the essential prop- 
erties of swans may be discovered. It is to be noted 
that any universal proposition asserts the non- 
existence of some class of things: that all swans 
are birds requires that there be no non-bird swans ; 
*See Appendix F. 


that all swans are white, asserts that the class of 
swans of different color is a class which has no 
members. But the proposition which is a priori 
does not assert any limitation of experience; it 
asserts only that whatever lacks some essential 
property, J5T, is not to be classified under some 
concept, A. That all swans must be birds, does 
not legislate out of existence any possible crea- 
ture. The empirical generalization, however, does 
require for its truth a limitation of nature and of 
experience : that all swans are white, excludes cer- 
tain conceivable creatures from existence. It is 
thus that the a priori proposition is assured with 
certainty in advance, while the empirical general- 
ization requires for its theoretical certitude a 
verification which extends to all reality. 

The empirical generalization is forever at the 
mercy of future experience, and hence probable 
only, while the a priori proposition is forever cer- 
tain* But as the above example points out, this 
does not represent any greater assurance about 
the content of future experience, or of nature, in 
the one case than in the other ; it represents only 
an intention of interpretation or classification 
which maintains a connection between two con- 
cepts regardless of experience in the one case but 
not in the other. Since the a priori in general is 
definitive and analytic, not synthetic, the case is 
the same for all a priori propositions. 

This particular example is trivial because the 


classification "swan" is not a very comprehensive 
one; its systematic interconnection with other 
classes and categories is relatively slight and un- 
important. But more impressive examples can be 
given to illustrate the same point that the a 
priori does not dictate to nature but concerns our 
interpretation of empirical facts. For example, 
the law of gravitation is a posteriori because, if 
it fails of verification, we shall still not abandon 
the concept "mass," or any of the other terms, 
but only the relation between them stated by the 
law. By contrast, geometry is a priori because if 
the sum of the angles of what is identified as a 
Euclidean triangle turn out to be other than 180, 
we shall condemn the experience as "mistake"; 
and if a sufficient number of such attempted veri- 
fications have, without exception, the same result, 
we shall abandon the Euclidean character of our 
space but not the meaning of "Euclidean trian- 
gle." I should suppose that the probability of 
Newton's laws and of those theorems of celestial 
mechanics which are purely geometrical is of the 
same order of magnitude. Certainly there is noth- 
ing in the a priori character of geometry to give 
us any superior assurance that experience will 
conform to it. In so far as certain principles op- 
erate as criteria of reality and apparent excep- 
tion to them condemns the experience as illusory, 
the a priori may seem to have another significance. 
But this is only because "nature" is itself a cote- 


gory the very fundamental and important cate- 
gory of the physical: what is extruded from it is 
still an absolutely given and un-get-overable fact 
of experience, requiring to be dealt with in some 
other way if we are to understand it at all. It is, 
in fact, easy to exaggerate the cleavage between 
the physical and the merely mental or psychologi- 
cal, such as the illusory, as one may observe by 
a serious consideration of the question whether 
mirages and mirror-images belong to nature or 
are merely mental. 

The facts which I should like here to emphasize 
are mainly two* In the first place, that no sub- 
stantive conception, determined a priori, is able 
to confine particular experiences within its con- 
ceptual embrace with absolute assurance ; that all 
identifications of objects and all material truth 
about future experience remains probable only. 
The supposition that any theory may secure for 
the a priori a different significance than this, is 
a delusion. The impossibility of it will become ap- 
parent if we remember two things: that experi- 
ence includes dream, illusion, and mistake as much 
as "the physical 55 ; and that no theory, even on its 
own showing, can attribute an a priori certainty 
which is not hypothetical to predications about the 
particular presented thing. In the second place, 
I would emphasize the fact that the whole body of 
our conceptual interpretations form a sort of 
hierarchy or pyramid with the most comprehen- 


sive, such as those of logic, at the top, and the 
least general, such as "swans," etc., at the bot- 
tom ; that with this complex system of interrelated 
concepts, we approach particular experiences and 
attempt to fit them, somewhere and somehow, into 
its preformed patterns. Persistent failure leads to 
readjustment; the applicability of certain con- 
cepts to experiences of some particular sort is 
abandoned, and some other conceptual pattern is 
brought forward for application. The higher up 
a concept stands in our pyramid, the more reluc- 
tant we are to disturb it, because the more radical 
and far-reaching the results will be if we abandon 
the application of it in some particular fashion. 
The decision that there are no such creatures as 
have been defined as "swans," would be unimpor- 
tant. The conclusion that there are no such things 
as Euclidean triangles, would be immensely dis- 
turbing. And if we should be forced to realize that 
nothing in experience possesses any stability 
that our principle, "Nothing can both be and not 
be/ 5 was merely a verbalism, applying to nothing 
more than momentarily that denouement would 
rock our world to its foundations. 

On the one hand, every concept, however un- 
important, gives rise to a formal truth exhibiting 
its structure, which it is beyond experience to in- 
validate and which in its own little way is a cri- 
terion of reality. The concept "swan" determines 
what is, and what is not, a real swan; though what 


is not a swan is, perhaps, some other kind of bird. 
And on the other hand, 710 concept or principle, 
however basic, can be guaranteed to bring lucidity 
and comprehension by being applied to particu- 
lar experiences in a predetermined way. Even the 
laws of logic prescribe only what is real thing, or 
properly determined event, and do not prevent 
those evanescent appearances and puzzling tran- 
sitions of experience which it baffles us to under- 
stand. On the one side, there is the Platonic heaven 
of our concepts, with the beautiful clarity of their 
patterned interrelations, and their absolute truth. 
On the other side there is the chaos of given ex- 
perience. The bringing of these two together is 
a matter of trial and error ; is that empirical and 
material truth which is never more than probable, 
and is subject to continual revision in the process 
of our learning. That kind of revision which 
means the abandonment of certain concepts as not 
truly applicable to certain areas of experience is 
more fundamental and important than the mere 
giving up of empirical generalizations previously 
held. But it is only a deeper-lying phase of that 
process which the progress of our understanding 
may necessitate. 

The truth of the a priori is formal only; but 
we cannot capture the truth of experience if we 
have no net to catch it in that is its immense 
importance. But so far as the validity of all ma- 
terial truth depends upon the predictability of 


particular experience, the problem of our knowl- 
edge of it is that of the validity of our prob- 
ability-judgments. That there may be no such 
valid knowledge because "there are no necessary 
connections of matters of f act," represents a prob- 
lem which is still to be met. 


The only knowledge a priori is purely analytic ; 
all empirical knowledge is probable only. In af- 
firming such a view, one assumes a heavy burden 
of proof ; the whole history of the theory of knowl- 
edge (unless we go back to Plato) seems to en- 
force the conclusion that such a conception must 
inevitably lead to skepticism. The presumption 
would seem to be that if the only general proposi- 
tions which are absolutely certain are of the con- 
ceptual, and if all empirical truth, including that 
about particular objects, is only probable, then 
there can be no genuine knowledge of nature at 
all; even genuine probability will be lacking, for 
probability itself must rest upon some antecedent 
certainty. More particularly, it may be felt that 
the knowledge of nature requires some ground of 
order in reality, or in the content of experience, 
which assures its consonance with our modes of 
conception ; that is, that there must be knowledge 
a priori of some "metaphysical" and synthetic 
truth, as contrasted with the merely logical truth 
of the analysis of concepts, in order to bridge the 
gap between abstract ideas in the mind and the 
reality presented in experience. Lacking this, 
knowledge of nature, since it is predictive, will 
find no basis for its validity. 



At once, it should be remarked that there is an 
absolute certainty of the empirical which has been 
recognized the immediate apprehension of the 
given. Such direct awareness is hot indubitable 
knowledge of an object, but the content of it is 
an absolutely given fact. This immediate presen- 
tation is our confrontation with reality and is 
requisite to the distinction of particular empirical 
truths from falsehood* Immediate qualia consti- 
tute the ultimate denotation in experience of our 
concepts, and the specific character of the given 
plays its indispensable part in any verification. 
It is difficult, if not impossible, to express the con- 
tent of the given without importing what is not 
given ; and our awareness of it has not been called 
tftf knowledge," because with respect to it there 
can be no error. Nevertheless, it functions as an 
absolute mv <rra> for the knowledge of nature, 

For the rest, as I shall hope to show, the con- 
ception that our knowledge of nature is a knowl- 
edge of probabilities, is the only one, compatible 
with demonstrable facts, which can save it from 
reduction in the end to mere "animal faith. 3 ' And 
furthermore, for the validity of empirical gen- 
eralizations as probable knowledge or more ac- 
curately, as knowledge of probabilities no a 
priori truth other than the merely analytic is re- 

It is true that, in order that the difficulties 
posed by skepticism may be met, it is essential 


that there be some knowledge which is more than 
probable, and that such knowledge should be per- 
tinent to nature and experience. But as has been 
pointed out in the last two chapters, this is se- 
cure: there is in all science, and in common-sense 
knowledge, an element which is absolute and cer- 
tain because it is a priori. The determination of 
the criteria of reality, in its various categories, 
and of principles of interpretation, antecedent to 
particular experiences, is purely analytic. Our 
concepts in general, without which no knowledge 
would be expressible and nothing in experience 
would be thinkable, give rise to such analytic and 
certain truth* As the matter is usually conceived, 
this knowledge which we find to be a priori would 
be included in what is meant by "knowledge of 
nature,*' since it delimits, for example, the physi- 
cal and prescribes basic laws which must be true 
of all physical reality. So far the point to be noted 
is, that there is a knowledge of nature which is 
more than probable because it is not merely em- 
pirical, or dependent on the content of the given* 
This is important because, as has been noted, one 
form which the skeptical difficulty takes is that 
empirical knowledge cannot be even probable un- 
less some knowledge is more than probable. The 
validity of probability- judgments rests upon an- 
tecedent general truths which must be certain. If 
all knowledge should be empirical and such prin- 
ciples therefore mere generalizations from ex* 


periences, then these principles would be only 
probable ; with the result that the knowledge which 
depends on them and is ordinarily called "proba- 
ble" would be only probably probable. Hence in 
ways which are obvious, any statement which we 
could make would require the qualification "prob- 
able," and knowledge would disappear in an in- 
finite regress of such qualification. It is the 
thorough-paced empiricism of a position such as 
Hume's which leads to this difficulty. Or to put it 
in another way, Hume's skepticism results, not 
from the absence of necessary connections of em- 
pirical particulars, but from failure to observe the 
ways in which the necessary connections of ideas 
are pertinent to the interpretation of the em- 
pirically given and hence are antecedent deter- 
minations of reality. 

Nevertheless, it is true that when our a priori 
conceptions are applied to given particulars, the 
truth to which they give rise is only hypotheti- 
cal, or if stated categorically, is probable only. 
Amongst universal propositions which refer to 
nature, we must distinguish between empirical 
generalizations which are synthetic such as the 
law of gravitation, for example and analytic 
principles which exhibit the consequences of our 
concepts, such as those of geometry* The former 
are probable only. The latter are a priori and cer- 
tain ; but their a priori certainty is either that of 
abstract conceptual systems, or when they are 


given denotation and application, it is hypotheti- 
cal, and when mention of the hypothesis is omitted, 
they are not certain but are merely probable. 
Thus a geometrical principle, when applied to a 
concrete presented object, is a priori and certain 
in the form, "If this plot of ground is triangular 
and our space is Euclidean, then the sum of these 
angles is 180 degrees.'* But when the hypothesis 
is dropped and we assert, "The sum of these angles 
is 180 degrees," the judgment is probable only, 
because there is no a priori and complete assur- 
ance that the concept "Euclidean triangle" is 
genuinely applicable to this plot. Likewise if the 
judgment is empirical but general such as "The 
sum of the angles of any plane triangle is 180 de- 
grees," it is probable only, because there is no 
complete assurance of the Euclidean character of 
space. Or if Euclidean conceptions are made defin- 
itive of space and hence criteria of "the spatial" 
as they might be then such an a priori de- 
termination must be accompanied by our pre- 
paredness to relegate any divergence of presented 
phenomena from Euclid to some other category 
to interpret them, for example, as physical re- 
fraction of light or as optical illusion* But the 
presence or absence of such divergent characters 
in some set of phenomena would not be determina- 
ble a priori or with absolute certainty. For exam- 
ple, there still could be no complete assurance 
that what are designated as "celestial triangles" 


bounded by light rays, were purely spatial phe- 
nomena, unaffected by some physical law of the 
bending of the rays. Hence any general empirical 
proposition about the set of actual phenomena, 
meant to be denoted by "celestial triangles/ 5 will 
still be probable only, because there will be no 
complete assurance that they have a right to the 
name and conform in all respects to what the a 
priori laws of space require of real triangles. 

As has been pointed out, every concept is cri- 
terion of some restricted kind of reality and, on 
the other hand, even basic or categorial concepts 
are not criteria of reality in general but only of 
reality within that category. Every presentation 
is an absolute fact; is the presentation of reality 
of some sort or other. But it does not follow that 
what is presented is classifiable in some particular 
category, such as the spatial or the physical, with- 
out mistake. Identification of what is presented 
as an object of a certain type, or a particular 
kind of reality, is an interpretation put upon the 
presentation, which is implicitly predictive and 
hence transcends the given and is subject to veri- 
fication or falsification by further possible expe- 
rience. If we know the properties of triangles or 
of space a priori, still the empirical judgment, 
"This is a Euclidean triangle," is no more than 
probable. Therefore the necessary connection be- 
tween "Euclidean triangle" and certain geomet- 
rical properties does not assure the geometrical 


truth about any particular presented object or 
any collection of such. 

Our subsumption of the given under concepts 
is, thus, always contingent upon future experi- 
ence, and the a priori knowledge of universal 
principles does not secure any a priori knowledge 
of empirical particulars. I think it will be clear 
that the connection between universal principles 
and empirical particulars has frequently, if not 
generally, been left a little vague, and open to the 
unwarranted inference that the certainty of the 
universal means an equal certainty about the par- 
ticulars, because the particular follows from the 
universal. It is true that "All triangles are thus 
and so" implies "This triangle is thus and so' 5 
(provided "this triangle" exists). But the point 
is, of course, that any presented this may not be a 
triangle but only an approximation to, or slight 
deformation of, a triangle, or something whose 
difference from the triangular is momentarily 
or for a thousand years incapable of detec- 

It has likewise been, most frequently, left a lit- 
tle vague just what is meant by "knowledge of 
nature" or "empirical knowledge," And this 
vagueness also is probably traceable to a failure 
of logical precision. Propositions of the general 

*We may remind ourselves that the Platonic distinction between 
knowledge which is a priori and of the idea, and opinion, which is 
probable and of the empirical, is based precisely upon the point that 
no sense-particular is exactly subsumable under any concept. 


form "All A Is J5" may have either of two mean- 
ings but not both at once.* They may mean (1) 
"The concept A includes or implies the concept 
jB" or (2) "The class, or collection, of A's is in- 
cluded in the class of JB's." In the first (the in- 
tensional) meaning, such a proposition is a priori 
true or a priori false ; there need be no appeal to 
experience to determine the implications of con- 
cepts. The second meaning is still not quite pre- 
cise until it is clear how membership in the class 
of A 9 s is to be determined. If it is determined by 
ideal conformity to the concept A, then obviously 
it will follow from the fact that this concept im- 
plies B 9 that the class of A*s so determined is con- 
tained amongst the ITs. But in that case, member- 
ship of any given particular in the class is al- 
ways subject to possible doubt. If, however, the 
term "A 9 s" is used denotatively to specify in ex- 
tension a certain group of particulars, then it 
does not follow from the fact that the concept A 
implies the concept J5, that a group of particulars 
called -4 5 s will indubitably have the character of 
B's. "Empirical knowledge" usually does and 
certainly ought to mean a knowledge of par- 
ticular things pointed out or otherwise deter- 
mined in extension. With this meaning, the em- 
pirical knowledge that a group of objects called 
"A's" will have the character of B's does not fol- 
low (as anything more than probable) from the 

*See Appendk F. 


a priori certainty that the concept A implies the 
concept B. The difference between the a priori, 
analytic, and intensional, on the one hand, and 
the empirical and extensional, on the other, is the 
difference between "If this is an A, then neces- 
sarily it is a jB" and "This is an A; therefore it 
is a J5." The former may be certain when the lat- 
ter is not. Hence the difficulty that a priori knowl- 
edge of universal truths does not lead to any cor- 
responding empirical knowledge which is indubi- 
table if this is supposed to be a difficulty is not 
peculiar to the conception here presented that a 
priori truth is analytic. Any theory whatever will 
have to meet this point unless the relation of uni- 
versal concepts and particular objects is somehow 

There is, to be sure, an important difference be- 
tween that knowledge of particulars which would 
be subsumable under a priori principles if only 
the applicability of the concepts in question were 
assured and a knowledge of particulars which is 
supported only by inductive generalization for 
example, between the inference from observed 
triangularity to other geometrical properties (if 
applied geometry is a priori) and the inference 
from observed physical character to gravitational 
behavior (laws of gravitation being inductive 
generalizations). In the one case, we know with 
certainty that if our identification and naming is 
not erroneous then the object will have certain 


further properties ; in the other we know only that 
if our identification is correct, it will probably 
have certain further properties. But in both cases, 
our knowledge of the presented thing is probable 
only because there is no complete certainty of its 
subsuraption under the concept which is in ques- 

Thus the compatibility of an a priori truth 
such as is here maintained with the thesis that em- 
pirical knowledge in general is no more than 
probable, turns upon the previous point that the 
knowledge of individual objects and of particular 
occurrences of objective properties always in- 
volves the application of a concept to something 
presented, and that this identification of the 
given as genuinely a case which falls under the 
concept is something which immediate experience 
does not make absolutely certain. Such identifica- 
tion is an interpretation which is essentially 
predictive and depends also upon a prior gen- 
eralization from experience. The identification is 
made on the basis of certain immediately pre- 
sented qualia which, in past experience, have 
proved more or less reliable clues to those further 
characters which are necessary to verification of 
the objective nature of the thing presented and 
to its valid subsumption under the concept used 
to name it. This interpretation reflects a judg- 
ment of the form, "That which presents the im- 
mediate qualia here given usually (and hence in 


a particular case, probably) has the objective 
character in question." 

This fact about the application of concepts to 
presented tilings emphasizes the general likeness 
of all empirical knowledge. Knowledge of indi- 
vidual objects, as much as of generalizations or 
laws, runs beyond what is given and asserts a cer- 
tain regularity or predictable interconnection be- 
tween experiences. Every objective judgment is 
such that it can be verified only by some progres- 
sion in experience. Since there is no indubitable 
knowledge of objects in direct awareness, the 
knowledge of things as much as the knowledge of 
laws is at the mercy of the future. Empirical 
knowledge depends on prediction, on an argu- 
ment from past to future, on the presence of some 
particular uniformity in experience ; and the gen- 
eral problem of its validity is the same which is 
posed by Hume's skepticism. How this validity 
can be assured without appeal to the dependence 
of the content of experience upon the mind, or to 
the limitation of experience in conformity to re- 
quirements of intelligibility, or to some other such 
metaphysical presumption this is, I should sup- 
pose, the outstanding problem which remains to 
be considered. 

At once, however, it is to be noted that this 
problem here assumes a form different from that 
in which it has usually been considered, in three 


In the first place, that part of Hume's skep- 
ticism which is concerned with necessary connec- 
tions questions the possibility of the knowledge of 
laws only and does not (explicitly at least) put in 
question the possibility of the identification or 
recognition of things. When it is seen that the 
validity of both these kinds of knowledge turns, 
for the most part, on the same considerations, the 
problem is considerably altered. At first glance 
it seems to be rendered even more difficult, be- 
cause the scope of the skeptical doubt is enlarged. 
But, as will appear later, this is not the case. It 
means that a world without law must likewise be a 
world without recognizable things. The recogni- 
tion of objects requires that same kind of order or 
reliable relatedness which law also requires. 

In a way, this means that the proof which Kant 
attempted in his deduction of the categories may 
be secured without the Kantian assumption that 
experience is limited by modes of intuition and 
fixed forms of thought. Because the deduction of 
the categories consist at bottom in this : that with- 
out the validity of categorial principles no ex- 
perience is possible. And a careful examination of 
Kant's argument reveals the fact that he uses the 
term "experience" to mean "objective experience, 9 * 
"valid experience," "experience of actual and 
identifiable objects," even though he does not 
make this quite explicit. He certainly does not 
mean that the categories are requisite to the ex- 


perience of a buzzing blooming confusion. In- 
deed, in some passages of the "subjective deduc- 
tion" the argument turns precisely upon the con- 
sideration that the only alternative to a cate- 
gorized and orderly experience is a meaningless 
flux of mere schwarmerei. Very likely the reader 
will incline to be harder upon the present attempt 
than upon Kant's famous argument, and to hold 
that it may not here be proved that knowledge is 
valid by showing that the only alternative is 
chaos. However, in advance of the argument, I 
think we may see that the question of the validity 
of empirical knowledge stands on a different foot- 
ing when we recognize that this is its only alterna- 
tive than when we suppose, as Hume apparently 
does, that we may take our world as a world of 
recognizable identifiable things while still doubt- 
ing the validity of all generalizations such as nat- 
ural law. 

The second remark which is in point is that 
nothing in the foregoing touches the very impor- 
tant fact that the principles of interpretation and 
classification and the criteria of the real are a 
priori and certain in advance of experience. This 
has an important bearing upon the problem of 
the validity of empirical knowledge in general, 
because it means that experience must, a priori* 
conform to certain principles in order to be per- 
tinent to any particular investigation or to the 
validity of any particular law of nature. Nothing 


is real in all categories ; everything is real in some 
category. A set of categories adequate to the un- 
derstanding of experience in general must meet 
this last requirement. It is not a priori certain that 
any given experience is validly interpretable in 
a particular category f or example, the physical. 
But we do know with certainty and a priori that 
if X is a physical thing, then it will conform to 
certain general principles which can be laid down 
in advance because they constitute criteria of the 
physical. When we study the sciences of physical 
reality, we have this a priori knowledge of prin- 
ciples to which the given must conform if it fall 
within the class of those phenomena which are sig- 
nificant in this connection. Thus we are provided, 
a priori, with a basal minimum of law within the 
field, as a sort of Archimedean point for all in- 
vestigation. This does not enable us to apply our 
basal principles which are a priori to some par- 
ticular given without the possibility of mistake. 
We may still be in error by confusing a subjec- 
tive and psychological phenomenon an after- 
image, for example with a physical thing. But it 
does enable us to be certain that nothing which 
concerns us in the study of physical nature can 
violate our fundamental principles. To fail to 
Conform is to be repudiated as not pertinent to 
our present study. 

Furthermore, if we should be possessed of an 
adequate.get of categories, then we may be certain 


a priori that whatever does not conform to the 
principles of a particular category will conform 
to the principles of some other which is coordi- 
nate. We play a sort of game of "animal, vege- 
table, or mineral" with experience, by which it 
will be impossible for it to get out of the net of 
our understanding, no matter what may be the 
content of it* 

The third important difference between what 
it is necessary here to establish and the problem 
as posed by Humian skepticism, is that it is 
the validity of empirical knowledge as probable 
judgment only which requires to be assured. If 
more than this is needed to save us from skep- 
ticism, then, once for all, there is no answer to 
the skeptic. The particular point of doubt, I 
should suppose, is whether probable knowledge 
and empirical generalizations can be valid if, as is 
here maintained, all necessary connections are 
logical only and do not limit the content of expe- 
rience. The validity of probability and of induc- 
tion is commonly supposed to rest upon some 
ground of order and connection beyond the 
merely logical some "uniformity of nature" 
which could conceivably be absent from our ex- 
perience. No such metaphysical presupposition 
would be compatible with the account of knowl- 
edge here given. For us, then, the validity of em- 
pirical knowledge turns upon the nature of the 
alternative when all assumption of more than 


merely logical necessity in nature, or of con- 
formity a priori to any order which could con- 
ceivably be absent, is dispensed with. 

For this, the first essential is an examination of 
what is essential to the validity of probable judg- 
ment from the purely logical point of view* It is, 
fortunately, unnecessary to enter upon a complete 
theory of inductive generalization and probabil- 
ity. Indeed, examination of this question could be 
dismissed altogether were it not that certain errors 
about the logical character of probable judgment 
have become entangled with the more fundamental 
question of the epistemological and metaphysical 
foundations of our knowledge of the probable. 
We must discuss these merely logical facts not to 
get them into the picture but to keep them out. 

"Probability" has many different meanings and 
the first requisite is to avoid certain verbal con- 
fusions which are possible. Ordinarily, we phrase 
the empirical generalization, or other statement 
which we know or should know to be probable 
only, as if it were absolutely certain. As economy 
of language, this is excusable and even unavoid- 
able. But if we thus state, for example, Newton's 
law of gravitation as absolute truth, we must not 
confuse what is stated with the judgment of any 
informed and intelligent person who makes the 
statement. The intent of the judgment is not the 
statement judged probable, but that it is prob- 
able. If in such a case we assert briefly that A is 


B 9 our judgment is, "It is probable that A is J3" 
or "that A is B is highly probable," or "A is B 
has a probability represented approximately by 
the fraction m/n" Now a common supposition 
seems to be that our knowledge of the law of 
gravitation is invalid if there are facts of nature 
which do not conform to the law. But if this is 
probable knowledge, it is a very simple and ob- 
vious fact that its validity does not require such 
conformity. The judgment "A is B is probable" 
does not require for its truth that A is B ; it re- 
quires only that this should be genuinely proba- 
ble. What the genuineness of a probability re- 
quires concerning the independent facts to which 
it relates, we need not, for the moment, inquire. 
But at least it is undeniable that the judgment, 
"A is B is probable," may be absolutely true when 
the judgment "A is J3" would be absolutely false. 
Unless this is the case, there is no real difference 
between probable and certain knowledge. No- 
body will contest this ; yet I think we may discern 
behind a good deal of the questioning about the 
validity of probable knowledge, the confused 
thought that if what we know is "A is B is proba- 
ble" but in fact A is not JB, then such knowledge 
is invalid. Certainly it is a question in what its 
validity consists, but equally certainly this does 
not consist simply in accord between what is prob- 
ably true and the objective fact. 

Another pertinent consideration has to do with 


the relation between the probable judgment and 
those facts which constitute the ground of it and 
hence stand to it in the relation of premises to 
conclusion* It has always been clear, of course, 
that it is impossible to tell whether a given judg- 
ment of probability is valid without examining the 
data on which the judgment is based. Every text 
on the elements of probability-theory contains 
some such illustration as the following: Let four 
hands of whist be dealt and each player inspect 
his hand. Let the quaesitum be the probability of 
four aces in one hand. This probability will be 
different (1) for A who is not a player and does 
not examine any hand (&) for a player 5, who 
observes that he holds no aces, and (3) for an- 
other player, C, who finds one or more aces in 
his hand but not four. If we represent the value 
of the probability for each by the letter designa- 
ing the person, then B> A > C, though for each it 
is the same objective stVte of affairs whose proba- 
bility is in question. In other words, a correctly 
estimated probability is relative to the data upon 
which it is based, or the premises from which it is 
drawn. This is frequently phrased by the some- 
what dubious formula, "Probability is relative to 
our ignorance." 

Thus probable knowledge is always relative to 
him who has it, in the sense that it depends upon 
whatever other relevant knowledge he may pos- 
sess. When some proposition or law is adjudged 


probable, there is always a tacit qualification 
which must be made explicit before the validity 
of the judgment can be assessed* For example, 
when the law of gravitation is declared to be 
probable, the real intention might be formulated 
as: "On the basis of what duly qualified persons 
who may fairly be presumed to know all the 
relevant facts which are available unite in re- 
porting, this law is probable. 55 The probability 
of the law is, very likely, different for me than 
for the scientist who knows the pertinent facts 
more directly. For him, the data are, in part at 
least, certain laboratory-tested facts. The rest of 
us having no such laboratory experience ac- 
cept it on authority ; which means that our judg- 
ment bears a less direct relation to the premised 
facts, and concerns the reliability of a proposi- 
tion about which expert investigators agree. For 
the scientist something is simply true which for 
me is probable because he reports it. Hence for 
me, the conclusion about the law represents a com- 
pounding of probabilities. (Strictly, of course, the 
same thing is true for the scientist also, though 
in lesser degree.) 

In this last respect, the illustration is typical 
of most probability-judgments. Nearly all the ac- 
cepted probabilities rest upon more complex evi- 
dence than the usual formulations suggest; what 
are accepted as premises are themselves not cer- 
tain but only highly probable. Thus our judg- 


ment, if made explicit, would take the form. The 
probability that A is B is a/b> because if P is Q, 
then the probability that A is B is m/n, and the 
probability of "P is Q 5? is p/q (where m/n X p/q 
= a/b). But this compound character of proba- 
able judgment offers no theoretical difficulty for 
their validity in general, provided only that the 
probability of the premises, when pushed back to 
what is more and more ultimate, somewhere comes 
to rest in something certain, and provided also 
that there are some valid principles of probability 
in general whether those commonly accepted or 
some others.* These two provisos, just stated, 
represent the prime requisites of the validity of 
probable judgment, concerning which there may 
be doubt. 

The validity of the judgment that A is proba- 
bly B does not, fits we have just seen, concern any 
direct relation between the judgment and the fact 
or non-fact, A is -B; it concerns the relation be- 
tween the judgment and whatever are the relevant 
data upon which it is based. These may be ver- 
bally quite remote; the immediate premises are, 
very likely, themselves only probable, and per- 
haps in turn based upon premises only probable. 
Unless this backward-leading chain comes to rest 

*StrjetIy r of course, there are theoretical difficulties about the com- 
pounding of probabilities, and in fact about almost every point in 
probability-theory. Except so far as they bear directly upon episte- 
moiogical problems, I have neglected these: short of interpolating 
here a complete theoretical analysis of probability, no other course 
is possible. 


finally in certainty, no probability- judgment can 
be valid at all. But if it does thus finally come to 
rest, the complexity of it is of no theoretical con- 
sequence. Such ultimate premises, however, must 
be actual given data for the individual who makes 
the judgment, hence the probability of a given 
formulation may vary from individual to indi- 
vidual, according to our individual knowledge of 
a relevant sort. 

All these facts are simple and fairly well rec- 
ognized; I hope I shall be pardoned for repeat- 
ing commonplaces in order to emphasize the fol- 
lowing obvious consequences of them: (1) In the 
only sense in which we can possibly suppose prob- 
able knowledge to exist at all, its validity is un- 
affected by the fact that it is subjective (that is, 
relative to the data of knowledge possessed by the 
individual, and very likely different for each) ; 
and () Probable knowledge may be valid in spite 
of the fact that what is judged probable may, in 
any given case or any number of cases, be false. 

There is a further important consequence of 
the relativity of probable judgment. Unlike de- 
ductive inference, in which the conclusion is as 
certain as the premises, the conclusion of a proba- 
bility-inference must retain its reference to the 
premises. The conclusion "A is probably J?" is 
elliptical ; what is validly meant is "On the prem- 
ises such and such, A is probably Jff. M This might 
easily be overlooked. One might say, "But since 


my premises are true, it is true without qualifica- 
tion that A is probably B." So phrased, the con- 
clusion "A is probably B" seems to refer directly 
to some objective fact. But it is just this over- 
sight which must be guarded against. As referring 
directly to objective fact, some new bit of evi- 
dence or the next moment's experience may com- 
pletely alter the probability may turn what was 
probable into something certainly true or cer- 
tainly false or something more probable or less 
probable than before. This is what happens when 
we say that the probability of something is in- 
creased or descreased. If, in the earlier illustra- 
tion, the man who has seen no hand of cards 
should be shown a hand with no aces, the probabil- 
ity of four aces in one hand "is increased," But 
the validity of his previous judgment is un- 
touched, and the fact if it be such that there 
are not four aces in one hand is unaffected. There 
is no such thing as the probability of four aces 
in one hand, or the probability of anything else. 
Given all the relevant data which there are to be 
known, everything is either certainly true or cer- 
tainly false. Given anything short of this, what 
the value of the probability is, depends upon what 
data are thus given. There are always various 
probabilities of the same qusesitum, on the basis 
of various data or different relevant knowledge. 
When the premises contain all the relevant knowl- 
edge which is available under certain well-defined 


conditions, then reference to "the probability" 
as in actuarial work, etc. has a recognizable sig- 
nificance ; it consists in tacit reference to this well- 
understood body of data. But clearly, this alters 
nothing in what has been said. 

A "poor evaluation" of the probability of any- 
thing may reflect ignorance of relevant data 
which "ought" to be known, or it may reflect logi- 
cal error in the relation of the probability con- 
clusion to its premises. In the former case, there 
will be moral or practical delinquency, perhaps, 
but the validity of the conclusion is unaffected. It 
is in the latter case only that any probability- 
judgment can be genuinely invalid. That the 
validity of probability-judgment is thus unaf- 
fected by ignorance and is affected only by logi- 
cal error, goes along with the fact that the con- 
clusion necessarily retains reference to its prem- 

The consequence of this which is most impor- 
tant is that the probable judgment, if valid, is 
true. There is no difference in the case of prob- 
ability-inference between validity and truth. What 
the judgment "A is probably JS" asserts is not 
that A is B or that any other objective state of 
affairs (except what the premises assert) holds 
good. It asserts that "A is B" has a certain prob- 
ability on the basis of certain data. If the data 
are actual the probability is "actual" ; if the data 
are merely hypothetical, the assigned probability 


shares this hypothetical character. But unless 
there has been logical error, the probable judg- 
ment is not only valid but absolutely true. There 
is no alternative to this account except that prob- 
ability has no kind of truth, no validity, and no 
meaning of any sort. 

Moreover, a probable judgment, once true, is 
always true. A probability cannot change, because 
probability has no meaning except by relation to 
its premises or ground. New data do not invali- 
date the previous judgment, because they consti- 
tute a new problem and mark a new probability. 
The probable judgment based upon specific data 
is not only eternally valid, if it is ever valid, but 
if it is valid, it is absolutely and eternally true. 

As we have seen, there are two types of empiri- 
cal generalizations with which we have to do: (1) 
those ordinarily so called, in which it is asserted 
that whatever may be validly named by some name 
has a certain further property, or properties, not 
implied by that name, and (2) those empirical 
generalizations of a subtler sort which underlie 
the naming of something presented. The first type 
include what are ordinarily called "natural laws" ; 
what they ostensibly assert is that wherever a cer- 
tain order is present in experience, a certain fur- 
ther order will accompany it. What those of the 
second type assert is that what presents a certain 
given appearance will exhibit in further experi- 
ence the order requisite to the applicability of a 
certain concept. 


For both types, the general character of the 
judgment is the prediction that something will 
hold of future experience because it has held in 
past experience. And in both cases, it is the valid- 
ity of this as a probability which requires to be 
established. One difference between these two 
needs to be remarked: "natural laws" must have 
held in all past experience and are predicted to 
hold universally, while this is not necessary for 
our interpretation of presentations. If I assert 
"This is a sweet apple," the nature of my judg- 
ment might be expressed: "What looks like this, 
under these conditions, will probably have the 
sweet apple taste, digestibility, etc." This judg- 
ment reflects my past experience of appearances 
like that now given. But it is neither plausible nor 
necessary that what looks like this should have 
turned out in all past cases to be a sweet apple. It 
is sufficient for probable judgment that this should 
have been so in a certain proportion of cases. 

The requirement of universality for natural law 
is, possibly, a bit artificial. What I mean is : there 
are any number of generalizations to be found in 
common sense and in practice which have the same 
general character as "laws" in other respects but 
which have not been universally substantiated in 
past experience and would not be regarded as in- 
validated by future exception : Potatoes are good 
food ; red-cedar shingles will last for twelve years ; 
a banker's advice about investments will be safe; 


the theatre-roof won't fall on your head. I think 
a little reflection will reveal that by far the greater 
part of life is guided by such generalizations 
which give rise to a probability in the particular 
case but are not without their exceptions. In fact, 
if laws of nature should have this character gen* 
erally, nobody would be much upset outside of 
academic circles. A generalization with very few 
exceptions is almost as good as one with none, as 
a basis for action. I am not trying to argue that 
there are no unexceptionable empirical generaliza- 
tions, nor to fudge the difference between cer- 
tainty and probability. But I would point out 
that, granting all the universal truth and all the 
certainty that the most ambitious theory has ever 
claimed, if it were not for that more lowly knowl- 
edge of probabilities based on generalities which 
have their known exceptions, we should most of 
us be dead within the week. A theory which ago- 
nizes endlessly about certain knowledge of nature 
and neglects the probable, represents a somewhat 
artificial interest. 

Thus we must recognize, alongside of those nat- 
ural laws which are based upon past experience 
without exceptions and are predicted universally, 
empirical generalizations admitting of possible or 
actual exception but nevertheless having a certain 
probability in the individual case. Let us call 
these last "statistical generalizations" since they 
are exhibited at their best when supported by 


statistical procedures. There are various theoreti- 
cal grounds quite apart from the practical con- 
siderations urged above on which it may be 
doubted whether such statistical generalizations 
and universally predicted empirical generaliza- 
tions, or "laws," can be distinguished in the end. 
But examination of these may be omitted, since 
decision of this issue is not of ultimate importance 
for us here. In any event, our knowledge of a gen- 
eralization is probable only, and the use of it de- 
pends simply upon its giving rise to a validly 
probable prediction in particular cases* 

It is obvious that all empirical knowledge even- 
tually goes back to knowledge of empirical par- 
ticulars. Generalizations have their ground in the 
coincidence of such particulars. Knowledge of the 
particular functions also as the basis of the ap- 
plicability of general principles which are not em- 
pirical but a priori. And knowledge of the par- 
ticular is rooted in immediate experience. The first 
apprehension, so to speak, is of given appear; 
ances, having a specific and later recognizable 
character, and of their continuity ; with further 
and equally specific experience. Coincidence of 
such progressions in immediacy give rise to habits 
of action, which may become explicit in general- 
izations of the form "What appears like this will 
turn out thus and so." Granted that such coin- 
cidence in experience can establish probability for 
the future, we have in the immediate awareness of 


the given that certainty which becomes the basis 
of a probable knowledge of the particular object 
or the occurrence of an objective property. 

The interpretation of the presentation is the 
application of a concept to it. The applicability 
of the concept requires, a priori, certain predicta- 
ble sequences in experience, continuous with the 
given. The application itself is hypothetical ; that 
the concept is genuinely applicable is not a priori 
but only probable. This probability is supported 
by a generalization from direct experience of the 
sort which has been pointed out ; a statistical gen- 
eralization to the effect that appearances like the 
given one, under circumstances like the present, 
are, in possible experience, continuous with such 
sequences as the applicability of the concept re- 
quires, in a certain proportion at least of cases. 

This probable knowledge of particulars be- 
comes in turn the basis for generalizations of the 
type more commonly recognized as such propo- 
sitions asserting a universal connection between 
what is denoted by some concept and a further 
character or property, not implied by that con- 
cept. Our knowledge of such generalizations rep- 
resents a compounding of probabilities, since its 
assumed premises the knowledge of particulars 
are only probable, and the passage from these 
premises to the generalization itself is inductive 
and represents a connection which is not certain 
but probable only. But this compound probability 


has as its premises the immediate certainty of the 
given data in the experience of particular in- 

At this point, the alert reader will douhtless in- 
quire if the validity of memory is not here as- 
sumed. The answer is, in brief, that this assump- 
tion is not necessary. Memory is a form of 
empirical knowledge, parallel in most respects to 
perception. As in perception, so in the case of 
memory, something is absolutely given the pres- 
ent recollection. And like perception, memory as 
a form of knowledge is an interpretation put upon 
this presentation; an interpretation, moreover, 
which in the particular case is verifiable, in those 
ways in which all knowledge of the past is subject 
to verification. Also memories, like perceptions, 
may be roughly divided into different types, hav- 
ing different degrees of reliability, which we are 
able to assess. This, plainly, is the only view con- 
sonant with common sense and with obvious char- 
acters of experience. Hence we must conclude that 
memory in general is probable knowledge, and 
that so far as other forms of knowledge are based 
on memory, the probability of such knowledge is 
compound, to a degree not previously noted. But 
that does not introduce any new theoretical dif- 

In fact, memory is in this respect like various 
other data which enter into the structure of our 
knowledge in general particularly the reports of 


other people. Such reports are more or less re- 
liable, as past verifications of such have attested. 
They enter into the body of data, upon which our 
further judgments are based, as more or less prob- 
able premises. So far as the experience support- 
ing an empirical generalization is thus vicarious, 
the probability of that generalization is compound 
to an extent which it would not be if the experi- 
ence were exclusively our own. Or to put it in an- 
other way, reports of others are a particular type 
of our own. experience, having a probability which 
reflects our past experience of such reports and 
of their relation to our further experience perti- 
nent to their truth. In various similar ways, our 
empirical knowledge is complex and remote from 
its bases in immediate experience. The probability 
attaching to it has a correspondingly compounded 

Some may feel that such an account makes our 
empirical knowledge so complex and, when ex- 
plicitly analyzed, so remote from its eventual 
grounds, that the kind of validity here assigned 
is little better than condemnation. But will not 
honest examination require the admission that our 
empirical judgments cure thus logically complex; 
that ordinary statements of them usually proceed 
by taking much for granted which is not abso- 
lutely certain. Such artificial simplification is ex- 
cusable, or even necessary, in the interest of the 
separation of problems, or for some other para- 


digmatic purpose. But it is in inexcusable in what 
purports to be an analysis of knowledge in gen- 
eral* If the truth should be complex and some- 
what disillusioning, it would still not be a merit 
to substitute for it some more dramatic and com- 
forting simplicity. 

There are, moreover, certain mitigating con- 
siderations. In the first place, the complexity of 
empirical judgment and its remoteness from com- 
pletely certain grounds, does not necessarily mean 
that its probability is diminished in like propor- 
tion* If the difference between the compounded 
probabilities and absolute certainty is slight, the 
eventual and resultant probability may likewise 
be very close to theoretical certainty. In the sec- 
ond place, the practical attitude which is ex- 
pressed in ordinary judgment as if it were based 
upon some certainty is, in fact, one which is quite 
compatible with the failure of what we say to 
prove true. Such failure would not prove devastat- 
ing to our attitude toward life as if an absolute 
truth had been destroyed. For example, if you ask 
me, I shall unhesitatingly assert, "This is a good 
fountain pen." But if next moment it refuses to 
write and thereafter can never be got to work 
properly, I shall not lie down and die because my 
knowledge is invalid and my universe has come 
apart. All my statement really means is that I 
have good enough grounds to think it highly prob- 
able which is, in fact, the case. If it should be 


pointed out that I am not even certain that this 
is the same pen I have used heretof ore, the experi- 
ence of which was my basis of judgment, I shall 
still not be disturbed in my practical attitude. I 
shall say, "Oh, well, what of it? Life is too short 
to bother about the difference between probabili- 
ties of that order and certainty." Which again is 
quite true. My unhesitating practical attitude is 
no mere "animal faith" ; it is quite in accord with 
the remote possibility that what I assert and act 
upon may not be true is logically remote from 
ultimate certain grounds and most complex. My 
attitude has a complete theoretical justification. 
It is precisely the supposition that knowledge re- 
quires absolutely certainty about the empirical, 
that closes the door to a theoretical justification 
of it. Such a supposition has no theoretical sup- 
port and no corroboration in our actual practical 
attitudes. There is just about enough, chance that 
our trusted generalizations may be false to make 
the pursuit of science pleasantly exciting. What 
a dull business life would be if everything we ven- 
tured to act upon should turn out true ! In a world 
in which knowledge, as some have portrayed it, 
would be valid, intelligence would be unnecessary, 
since habit would be a universally safe guide. 

If now the analysis of our empirical knowledge 
can be supposed to be covered, let us turn to its 
general character. It is that of a probability- 
judgment which, when explicitly stated, affirms 


that something has a certain degree of probability 
on the basis of premises which are, eventually, di- 
rect individual experience. The validity of such 
knowledge does not, of course, require that we 
should explicate all its complexity, or even that 
we should be able to. The man whose shrewd but 
untutored logical sense prevents his believing 
what logic would condemn, makes a valid judg- 
ment, whether he could provide the logical anal- 
ysis of it or not. If our empirical judgments in- 
clude only what a just logic would validate, they 
are sound. And in so far as we do not offend 
against logic, but hold to our empirical knowledge 
as probable in a degree which it truly is, the con- 
tent of our knowledge is true. Experience next 
moment may destroy the generalization judged 
probable. But it will remain forever true that it 
was probable on the grounds from which we made 
our judgment. And that is all that any prob- 
ability-judgment can validly mean. 

Since valid empirical knowledge means only 
such probability, on grounds which genuinely es- 
tablish it, and since any other than empirical 
knowledge is a priori, we have the important con- 
sequence that just m so "far as we are rational, 
what we believe is absolutely and eternally true. 
What rational men entertain as highly probable 
may largely alter with the passage of time. Em- 
pirical generalizations, as usually phrased, may 
be overturned and others take their place* The 


growth of science may repudiate as "false 35 what, 
in its previous stages, was held "true." But a just 
appreciation of the nature of such knowledge as 
only probable has this consequence : let our igno- 
rance be however large, our experience however 
circumscribed, we need believe nothing false, ex- 
cept as we fail to be rational and believe without 
valid grounds. Such avoidance of the unwarranted 
will not condemn us to sheer ignorance; we may 
at every stage possess a generous body of gen- 
eralizations which, correctly assessed, are valid 
and are useful guides to practise. Indeed, will not 
a survey of the history of human thought compel 
the conclusion that only such a conception as this 
can save the reasonable-minded man from repu- 
diating the attempt at scientific knowledge as 

However, it may be said that what the defense 
of knowledge requires is not a justification of it 
as a logically valid judgment, corroborated by 
reference to past experience but doomed perhaps 
to repudiation in the light of the future* Empiri- 
cal knowledge is essentially predictive and its re- 
lation to the future is of the essence of its validity. 
There is no attempt to escape this point. I have 
so far spoken as if, in general, the argument from 
past to future is valid; as if some ground of in- 
duction is secure. The fundamental doubts which 
may be entertained of this will be the topic of the 
concluding chapter. 


Two points may be noted here: first, a "law" 
having a high degree of probability may have to 
be abandoned in the light of future experience as 
not universal, and may yet remain a "rule of 
thumb" or statistical generalization which is true 
in the great proportion of cases and hence gives 
rise to a still valid probability in the particular 
case. The practical use made of laws which are 
superseded may, and often does, stand as still 
justified. The "laws" of the ancients are, often- 
times, such as would still be useful guides to ac- 
tion if we had no better. And second, we need to 
ask just what is necessary for the justification of 
probability-judgment as a basis for practical ac- 
tion. That what is probable must always be true, 
is an obviously impossible answer to this question. 
It is even impossible to demand that whatever is 
probable must in every instance be true in the 
majority of cases in spite of some theories to the 
contrary. A probability may genuinely be valid 
in some instances even though beyond a certain 
point no case should be found in accordance with 
it. I think reflection will reveal that what is req- 
uisite to its justification as a practical attitude 
is that action in accordance with probabilities 
must in general be more successful than action 
which ignores them. In other words, it is essential 
that the world be such that probabilities .in gen- 
eral are justified by the future that the world 
is "orderly"; that there are certain stabilities ex- 


tending through the past and future, and that 
the attitude which is based on past coincidence 
will in general be safer for the future than a dif- 
ferent one. But this question is that same one con- 
cerning the existence of some basis for induction, 
which the next chapter will discuss. 


Since empirical knowledge is exclusively a 
knowledge of probabilities, the validity of it in 
general depends upon the validity of induction 
and probability-judgment. The preceding chap- 
ter has been written as if some principles of such 
inference may be presumed as valid. That pre- 
sumption, however, requires justification, particu- 
larly since the grounds upon which it is often 
supposed to rest have here been repudiated. Let 
us phrase the issue as sharply as possible: Con- 
cepts are of the mind* All knowledge is in terms 
of concepts, and the possibility of it depends upon 
their applicability to experience* The application 
of a concept requires always a certain orderly 
sequence in experience. But the content of ex- 
perience is independent of the mind; that order 
is discoverable in possible experience, cannot be 
dictated by the knower. If generalization is to be 
possible if concepts are to be applicable in dis- 
tinguishable classes of cases, and if the connection 
of concepts in such generalizations is to find its 
application in reality then the givenness of cer- 
tain qualia, or complexes of such, must be a clue 
to expected sequences, and the occurrence of such 



sequences in the past must be a ground of their 
valid prediction for the future. 

However, it is not requisite that such expecta- 
tion and predication should be certain. Knowl- 
edge of particular objects is never beyond the 
possibility of mistaken apprehension ; and empiri- 
cal generalizations are, theoretically, never more 
than highly probable* What is requisite, then, in 
order that empirical knowledge should be valid, is 
that this connection of given qualia with expected 
sequences, and the connection between the se- 
quences prescribed by one concept and that which 
is essential to some other, should be genuinely prob- 
able. In general there must be the possibility of 
arguing from past to future ; not with certainty, 
but with probability. 

Concerning probability-judgments, it has been 
pointed out, first, that their validity does not re- 
quire that what is judged probable should be 
true not even that the particular generalization 
should be true of any specifiable proportion of 
actual cases* Second, the probability- judgment is 
relative to the pertinent knowledge of him who 
makes it ; and this relativity is no bar to its valid- 
ity. Third, this relativity of the probability- judg- 
ment to its premises means that its validity, in the 
particular instance, consists in a certain relation 
between the conclusion and its ground ; and means 
also that if it is valid it is true, absolutely and 
eternally. These three characters of probability- 


judgments rest upon obvious facts which cannot 
be denied without destroying the distinction be- 
tween probability and certainty. It is an im- 
mediate consequence of them that a particular 
empirical judgment, if it represents a probability- 
inference justly drawn from its grounds, is abso- 
lutely true knowledge* The one remaining ques- 
tion is whether there are any valid principles of 
such inference according to which a particular 
empirical judgment may be justly drawn. If prob- 
ability-judgment m general may be valid then 
there is no further ground of doubt that empirical 
judgments which are rational are true. 

For the ideal completion of the argument, pres- 
entation and detailed examination of the princi- 
ples of probability-inference and induction as 
well as of our categorial concepts would be in 
order. But the reader will not ask for that in the 
present book another of at least equal length 
would be required. The particular ground of pos- 
sible doubt is obvious enough: the applicability of 
concepts and the argument from past to future, 
require the presence of some order and uniform- 
ity. In an experience whose content is independent 
of the mind, it may be thought that such order 
could conceivably be lacking, and that the pre- 
sumption of it is, therefore, dogmatic and without 
foundation. Pointing out that the validity of 
probability consists in a relation between the con- 
clusion and its ground, and that the truth of 


what is judged probable is not directly relevant , 
does nothing to meet this present point. Precisely 
what is in question is whether a judgment which 
is in this sense subjective, confronting an experi- 
ence which is independent, can be meaningfully 
relevant to the constitution of reality. 

Since the applicability of concepts (or recog- 
nition of things by their appearances) as well as 
the validity of generalizations, is in question, the 
issue concerns the intelligibility of experience as 
well as the possibility of empirical knowledge. 
These two turn, for the most part, upon the same 

The conclusion of which I shall hope to con- 
vince the reader is that no assumption of any- 
thing which could conceivably be false is neces- 
sary ; that no sort of experience which the wildest 
imagination could conjure up could fail to afford 
a basis for intelligibility and probable judgment. 
The contrary assumption has frequently been due 
to the false conception that is certainty of appre- 
hension and certainty of generalization which 
must be provided for. And some of those same con- 
fusions which we have found surrounding previous 
issues are involved here also. 

I recognize that my burden of proof in this 
matter is a heavy one. Belief in something meant 
by the "uniformity of nature" is 5 1 think, as natu- 
ral to us as belief in an absolute up and down, and 
is supported by many habits of thought which are 


fundamental and pervasive. And so far as I know, 
it has never up till now been questioned, except 
by those who willingly faced a skeptical alterna- 
tive. In a sense, this belief is not to be questioned 
here, but rather whether it has any alternative at 
all; the precise problem is, perhaps, just what is 
involved in the necessary "uniformity." In this 
difficult situation, instead of proceeding directly 
to the center of the problem, I wish to begin with 
a variety of more peripheral considerations. Per- 
haps if a sufficient number of external buttresses 
are removed, the false conception will fall of its 
own weight. 

Two points which are immediately relevant to 
the question of order in reality can be brought 
forward from what precedes; first, that reality 
and the content of experience are not directly [ 
synonymous, and second, that our categories are 
so divided that always we play a sort of game 
of "animal, vegetable, or mineral" with the given. 
It is reality, not experience, which must be or- 
derly. Failure of a certain type of order is the 
criterion which excludes the given from reality 
(of a certain type). Thus so far as any one cate- 
gory is in question, our method of understanding 
experience is to segregate, as "reality," that part 
which is orderly in the required fashion ; the rest 
is understood by being labelled "unreal." With 
this in mind, our rational demand that reality 
shall be orderly somewhat reminds one of the silly 


old story we used to tell as boys of the man who 
made a list of those he could whip : when a neigh- 
bor whose name was included belligerently af- 
firmed, "You can't whip me," the mater of the 
list replied, "All right; then I'll just rub your 
name off." Experience has not much chance to 
thwart our demand for order when the failure to 
be orderly in certain ways merely results in its 
being rubbed off the list of that which it is de- 
manded shall be thus orderly. 

To be sure, what is excluded from one category 
must be brought under some other if it is to be 
intelligible. But any set of coordinate categories 
is simply a method for exhausting the possibili- 
ties. The "unreal" is a temporary pigeon-hole for 
what requires to be sorted or analyzed m some 
further fashion. The unsatisfactoriness of such a 
scrap-basket category merely reflects this desira- 
bility of a further understanding of its content. 
But to be able to classify what is presented as 
"unreal" or "illusion," though it may represent 
only superficial understanding, means neverthe- 
less a very important understanding, precisely be- 
cause it means that this content of experience is 
not relevant in the present connection, that it can- 
not figure as a negative instance of an empirical 
generalization and so on. So far, then, the reality 
presented to us in experience is certain to be in- 
telligible and orderly, because the failure to make 
a certain kind of sense merely results in its being 


relegated to the box we keep for pi. The only 
question is, how much of experience will be reality, 
and how much illusion. It will be obvious that this 
depends, in part, at least, upon the intellectual 
ingenuity of the knower his power, when some 
expected order fails, to discover some other which 
is definite to like degree. It is further clear that 
to the question, "How much of given experience 
will be illusory?" there can he no a priori answer. 
In this connection it is well to remark that our 
understanding of the given is always a matter of 
degree, and that the order demanded of reality 
is, similarly, more or less specific. If there is any 
sense in which the real must be "through and 
through" orderly, at least such through and 
through order is an ideal correlated with that 
"complete understanding 59 which is impossible to 
any but an absolute mind* And there must be some 
other sense in which predictable order is not de- 
manded of the real. The reasons for this will bear 
a little investigation. 

I find this morning on my study window a num- 
ber of apparently random grayish smudges. The 
explanation promptly comes to mind; small chil- 
dren played here yesterday. Just this pattern of 
smudges probably never occurred before and 
never will again. But this failure to exhibit any 
definitely anticipatable arrangement excites no 
surprise ; it is just what one expects of children's 
finger marks. This superficiality of our demand 


for order reflects, in part, the lack of any further 
cognitive problem. If my choicest possessions were 
all missing this morning, I should ask for some- 
thing more. The detective would come in and, 
starting from a system which exhausts in certain 
ways the possibilities of finger-marks, would seek 
to establish a correlation identifying a thief. But 
even so, these finger marks would still fail to be 
uniform with anything else in some ways in which 
order could exist and might be expected. The 
phrase "through and through" as applied to the 
uniformity or order demanded of reality is rather 

It also suggests itself that, in the process of our 
learning the nature of the real, what we do is to 
look for some order of a certain general type and, 
if we do not find that, to look for some other. The 
first attempt at uniformity may be, for example, 
that sparks fly upward, water runs down hill and 
"Everything seeks its natural level." Balked in 
this, a second attempt is, "Bodies fall in propor- 
tion to their weight." Finally we have z>=# *. The 
point here is not simply that one attempted gen- 
eralization having failed, we seek another. It is 
that the type of uniformity first sought a cor- 
relation-Jbetween gravitational behavior and physi- 
cal kinds does not exist. Nor is there any cor- 
relation with weight. The correlation found be- 
tween velocity and time is of an entirely different 
sort. Not order "through and through" but some 


order is what is requisite to intelligibility. And in 
the light of the nature of our learning process, 
the dictum "There must be some order in any 
given area of reality" takes on the character of 
a regulative principle, not particularly different 
in significance from, "If at first you don't suc- 
ceed; try, try again." More explicitly and ac- 
curately, the situation may be stated thus : A cer- 
tain minimal order Is prescribed a priori in the 
recognition of the reaL It is a regulative maxim 
of reason to seek further uniformities which may 
be stated in principles finally of maximal compre- 
hensiveness and simplicity. But there neither is 
nor can be any prescription of the specific type of 
uniformity or correlation which is demanded in 
this interest of further intelligibility. Moreover, 
the particular kind of order discoverable in one 
segment of the real may be definitely absent in 
some other segment in which it might with equal 
reasonableness be anticipated. 

The situation is entirely comparable if we turn 
from the kind of uniformity necessary for gen- 
eralization to the kind which is essential for the 
recognition of objects. A certain minimal uni- 
formity is prescribed by the categorial classifica- 
tion as reality of a certain type. Purther uni- 
formity may be sought for the purpose of further 
classification. But the manner in which such fur- 
ther uniformity shall be f oupd is not determinable 
unless by some scheme of subordinate cate- 


gories, exhausting the possibilities in a particular 
wa y i n advance of familiarity with the area of 
reality which is in question. Nor can it rationally 
be "demanded" that any particular degree or kind 
of such uniformity shall be always exhibited by 
experience which is intelligible. What the recog- 
nition of objects requires is some correlation be- 
tween given appearance and that sequence with 
further experience which is requisite to identi- 
fication and to discrimination of veridical from 
mistaken apprehension. But it cannot be re- 
quired for the intelligibility of experience that 
within the limits of the given there must be that 
which affords the basis of such uniform correla- 
tion. If within every empirical content merely as 
given there were that which possessed absolute uni- 
form correlation with that further sequence which 
is essential to a correct apprehension, then illusion 
and mistake would be possible only to the inex- 
perienced and the fool. It is true that only the 
irrational need be, in the strictest sense, deceived 
by appearances; but this is because the rational 
must realize that mistaken apprehension is al- 
ways possible and identification in general are 
only probable. The possibilities of sequence be- 
tween the appearance and further experience 
must be at least dual wherever the experienced 
and intelligent observer finds even the slightest 
possibility of mistaken apprehension. Obviously 
it is not the case that every given, quale or every 


complex of such has some uniform correlation 
with something else, sufficient to render it intel- 
ligible in the manner knowledge seeks, else we 
should all of us stand convicted of stupidity to a 
degree which is quite implausible. 

There is another way in which it can be made 
evident that we cannot require that experience, 
in order to be intelligible, must be such that 
something in the given content of any experi- 
ence is uniformly followed by something else in 
further experience. If this were the case, then 
within every experience merely as given must be 
something which determines absolutely this fur- 
ther experience which is predictable from it. This 
further experience, in turn, would likewise dic- 
tate some future experience, and so on. Thus any 
given experience would uniquely determine the 
course of future experience, or at least of some 
endless chain of further experiences. It may seem 
that this is precisely what is required for knowl- 
edge and prediction. But it is not: in a world so 
constituted whatever could be learned would not 
be worth knowing, because nothing could be done 
about it. It is worth learning that hot stoves burn 
precisely because the feeling of radiated heat does 
not inevitably determine that further sequence 
which we first investigated to our cost, but only 
determines it under certain further conditions 
which we take care not to allow. There is a uni- 
formity in reality hot things burn which is 


definable as a uniformity of possible experience. 
But the antecedent in terms of which the conse- 
quent, being burned, is determined, is something 
much more complex than what is confined within 
the given experience of radiated heat. Otherwise 
we should fatally go on to be burned every time 
we felt it, just as we did at first. Now some things, 
which are predictable, are doubtless unavoidable 
in experience. But even in these cases ; it does not 
necessarily follow that they are determined con- 
sequents of given experience. It is possible and 
much more likely that here too the antecedent 
of which they are uniform consequents is some- 
thing much larger and more complex than any 
previous given experience. Often our inability to 
avoid them is due to our ignorance of the fur- 
ther conditions, beyond our experience, upon 
which these consequents follow. If we knew 
enough, we could still avoid them, and that is the 
agony of our situation. And often also if not 
universally where some future experience is be- 
yond our power to avoid, and would still be be- 
yond our power even if our knowledge should be 
greater, it is nevertheless the case that the deter- 
mining antecedent is not wholly within our expe- 
rience but contains further conditions outside it 
but also outside our capacity to change. If it were 
not for these further conditions, we might still 
have precisely this experience and yet escape the 
denouement. Quite clearly then, knowledge does 


not require that kind of uniformity which would 
mean that something given in experience is uni- 
formly followed by something further in experi- 
ence* If this were so, then life so far as we could 
know and understand would be merely a fatal 
unfolding of the inevitable, and our knowledge 
would be a worthless revelation of that fate. Even 
those who read life and reality in terms of such 
inevitable unfolding, do not condemn us to foresee 
it step by step merely through intelligent inspec- 
tion of our given experience. As has been pointed 
out, those predictions which are the primary con- 
stituents of our useful knowledge of nature are 
of the form: Since X is given, if condition F 
should be supplied, then Z would accrue. Where 
F is a condition which I myself fulfill, or refrain 
from fulfilling, my knowledge serves to guide my 
action to desired ends. The sweetness of the ap- 
ple, the hotness of the stove, etc., are known by 
means of such truth of hypothetical propositions : 
this round, ruddy somewhat being given, if I 
should bite, it would taste sweet ; this visual pres- 
entation and feeling of warmth, being given, if 
I should touch, I should be burned. As has been 
pointed out, if I could do nothing about experi- 
ence, then since such, hypotheticals would be mean- 
ingless, reality would be no thicker than an in- 
evitable stream of consciousness that is, I should 
not confront reality but at most only a fatally 
determined life. Knowledge of reality serves for 


the control of experience: without the possibility 
of control., not only would knowledge be worthless ; 
there would be for us no reality to know. Both 
the usefulness of knowledge and the meaningful- 
ness of reality require that the uniformity appre- 
hended by knowledge should not be such that the 
determining antecedent is completely contained in 
the eixperience now fixed by being given. It is re- 
quired for the significance of knowledge and the 
real that uniformities be specifiable as probable 
at least in terms of possible experience. But be- 
tween possible and actual experience is the whole 
of that which differentiates reality from mere im- 
mediacy. Whatever the uniformity of reality re- 
quired for knowledge may mean, it cannot mean 
a fixed and uniform sequence in which given ex- 
perience is one complete term. 

It is further to the point that the whole effort 
of intelligence and those habits of action which 
are presumably its genetic antecedents, are bent 
to the apprehension of whatever in the given may 
constitute a distinguishing mark and serve the 
purpose of prediction. The instant mental reac- 
tion to experience, the manner in which we ap- 
proach it and the way in which we abstract from 
it the presentation of objects, reflect millenia of 
nature's work to the end that we may grasp what- 
ever in experience is clue to some uniformity of 
the sort which intelligibility requires. Such char- 
acters of the given arrest the attention and are 


for us something, while that which does not thus 
possess meaning slides off the surface of the mind ; 
it requires a reversal of all that is natural and 
habitual for us to catch it, and there is no word 
by which it may be held. Hence it is almost in- 
evitable that the extent to which actually given 
experience is uniform or contains clues to uni- 
formity, should be exaggerated to our casual in- 

There is another consideration which should 
be added to this. It is a rational demand, of at 
least as good standing as the demand for uni- 
formity, that every individual object shall be 
unique. Such uniqueness cannot reside in the 
given appearance of the thing: the number of 
sensory qualia we can distinguish is finite, and the 
number of combinations and arrangements pos- 
sible within the mental field of a single experi- 
ence, though large, is totally inadequate to 
uniqueness. This uniqueness can only reside in the 
further specifiability of the object in possible ex- 
perience continuous with the appearance of it; 
that is, in something which is true about the ob- 
ject, though it is not at this moment apparent, 
and something which, by appropriately directed 
investigation, we could learn. Hence this demand 
for uniqueness requires that every recognizable 
appearance must be correlated with some further 
verifiable specification of the object, in a way 
which is different in at least some one respect from 


each and every other. When the multiplicity of 
objects, actual and conceivable, is contemplated, 
it becomes clear that the extent to which experi- 
ence is thus required to be non-uniform is in- 
definitely large. 

I do not here defend the theoretical consistency 
of such conception of uniqueness, and I do not 
accuse any one of holding to it in this form; 
any more than I would attribute the theory of 
"through and through" uniformity to anybody in 
particular. My object is to let one shibboleth fight 
another, to the end that certain vague and unex- 
amined modes of thought may be dragged into the 
light and certain superstitious and equally vague 
beliefs about experience may be destroyed. The 
most that can reasonably be believed is that ex- 
perience when caught in the net of our categories, 
will always afford some clue to an actually existent 
further uniformity of some sort. Identifiable or 
recognizable appearances in their continuity with 
further experience, must be at least as much non- 
uniform as uniform. 

A further simple illustration of the way in 
which non-uniformity may be remarked and yet 
have no significance for knowledge, may be of ser- 
vice* There is a popular superstition that no two 
snow crystals ever exhibit precisely the same pat- 
tern. I have no idea what warrant there is for it, 
and perhaps the reader has not either. We can 
see that the number of possibilities of such crys- 


talline forms is indefinitely large, but within the 
limits of size and of the smallest discernible ele- 
ment, it would be finite. Suppose that neither gen- 
eral uniformity nor uniqueness should obtain. 
Nothing in this would be baffling to our recogni- 
tion of snow or our knowledge about it* Whether 
this illustration really serves or not, at least na- 
ture is full of such frost-patterns, the leaves of 
trees, the structure of growing plants, etc. Yet 
those relatively slight and superficial ways in 
which two oak-leaves or two frost-patterns are 
alike, suffice for our recognition of them and our 
generalizations about them. A non-uniformity, to 
be significant for knowledge, must be very specific 
it must be a negative instance of our predictive 
recognition or attempted generalization. And even 
when we are thus thwarted, we simply give up our 
previously held specific mark or law, and proceed 
to understand that kind of object in some other 

Quite frequently that order which, intelligibility 
and law require has to be sought at some one level 
and escapes us at other levels. A general illustra- 
tion is the whole body of those phenomena recog- 
nized by science in which macroscopic uniformity 
is superimposed upon microscopic multifarious- 
ness. That law at the macroscopic level may even 
be based upon an assumed randomness of the 
microscopic, is instanced by the kinetic theory of 
gases. As is well known, it may be held that law 


in general has this character. Whether this the- 
ory of law as statistical generalization of random 
distribution at some lower level is defensible or 
not, at least it is of great importance that the ab- 
sence of observable order which is called "chance" 
itself becomes a kind of uniformity and may give 
rise to law. 

This type of consideration is, it seems to me, an 
observation about our modes of dealing with phe- 
nomena rather than about nature or experience. 
There are other types of illustration also of the 
fact that certain kinds of order are imposed where 
none is directly observable. The outstanding ex- 
ample is, of course, arithmetical order, which pre- 
sumes nothing more than some kind of identifia- 
bility. I do not refer here to the counting of as- 
semblages; important though that may be, still 
it is less important than the serial arrangements 
imposed upon intensive (that is, qualitative) dif- 
ferences* At least I believe that this is what ex- 
amination of our categories would reveal. The 
multiple types of such "arrangements" of what is 
in no sense arranged in nature or experience 
either before or after will need no comment ; nor 
will the fact that such imposed order quite gen- 
erally provides the basis in terms of which other 
types of order are expressed. 

At the risk of going beyond what could be 
clear without detailed examination of our cate- 
gories, I would draw attention to two character- 


istics of such, imposed order.* In the first place, it 
allows an infinite multiplicity to be brought un- 
der a finite simplicity of rule. It may seem trivial 
to remark that ten characters and a rule of add- 
ing 1 give us command of infinity, but without 
that, or something similar, there would be no point 
in counting though there might be in serial ar- 
rangement. And second, correlation between nu- 
merical order and an imposed serial arrangement 
of qualities extends the power of this type of or- 
der to what is in no sense countable. Such imposed 
order, which demands no sort of uniformity of 
nature beyond the persistence of identifiable char- 
acters, is at least a prime constituent of intel- 
ligibility, though without some sort of correlation 
between it and that further kind of order which 
means determinable sequence of experience, it 
would not possess the significance which it does. 

This type of cor r elation t can be illustrated by 
the case of color. The distinguishable color-qualia 
are not unlimited in number but are conf usingly 
numerous, and also they are subject to the acci- 
dent of indiscernible difference; B may be indis- 
tinguishable both from A and from B, though A 
and C are recognizedly different. We make a vir- 
tue of this difficulty by translating the AB-C 
qualia as a continuous series. The manner in 
which this procedure is systematically carried out, 

*It will occur to the reader that, according to Cassirer, such dimen- 
sional and serial arrangement is the universal type of conceptual order. 


in the color-pyramid, needs no exposition here. It 
should be noted that this arrangement would per- 
mit of algebraic treatment, quite independent of 
any correlation between color and harmonic mo- 
tions, which is a further order, of the type of dis- 
covered generalization. 

Color also illustrates another common method 
of achieving simplicity; that is, by dividing a 
whole field of qualia into classes by the use of 
names with a qualitative range of denotation. 
"Red" or "blue" represents 110 single quale, but 
instead a considerable variety of such. That the 
mind could hardly make a beginning of bringing 
order into given experience without this device, 
should be evident. It is made use of wherever it is 
the case that no imaginable instance can com- 
pletely contain or illustrate the essential denota- 
tion of the concept as there can be no image of 
triangle in general or dog in general. This as- 
signment to a name of a range of denotation 
should be sharply distinguished from that ab- 
straction of the essential and ignoring of other 
characters which is represented by many theories 
as the universal basis of general names. Such 
theories are, of course, inadequate in several 

( *The real basis of classification of the kind in question is, of course, 
similarity. Similarity is of two types, partial identity and resemblance 
proper. A spatial or temporal whole, like a contour or a melody, may 
be divisible into parts some of which may be qualitatively identical. 
But similar color-qualities are an instance of the other type. Resem- 
blance means the possibility of confusion. It is apprehended by con- 


The substitution of simple classification based 
on resemblance for an indeterminately large num- 
ber of distinguished qualia, is of considerable im- 
portance in connection with probability- judg- 
ments. For instance, there is a finite probability 
that this book will be bound in red, because it 
must be bound in red, orange, yellow, green, blue, 
purple, black, white, or gray, if it is bound at all. 
The determinable probability arises through the 
fact (among others) that, within the universe of 
discourse in question, the possibilities are ex* 
hausted by a definite number of categories. 

It is evident that the last few pages have 
touched superficially upon a variety of topics 
relative to the categories, each of which is worthy 
of extended examination. But as has been said, 
these matters are not of central importance to the 
main issue. The points which I hope have been 
suggested, if not established, are the following: 
Reality is more orderly than experience, because 
reality is experience categorized. Lack of certain 
types of anticipated order leads to repudiation 
of the given content as "unreal.' 5 The "unreal" 
must be capable of being understood in some 
other way but understanding is always a mat- 

scious or unconscious recognition of this possibility. A man "looks 
like** my brother if, at a quick glance or at a distance, I might mis- 
take him for my brother. Things are more or less similar according 
as optimimum conditions must be more or less nearly approached 
for distinction to be made. Partial identity may also be included 
under this rubric. It should be noted that recognition of similarity is 
a kind of latent generalization. 


ter of degree, and the designation "unreal" ordi- 
narily marks that type of superficial understand- 
ing which characterizes the situation in which it is 
important only that a particular given experi- 
ence be excluded from the field or what is relevant 
to the problem in hand. All generalizations are 
based upon reality, not upon uncategorized expe- 
rience. Neither reality, or nature, nor experience 
is orderly in the sense that what is presented may 
be taken in any way we please and found to ex- 
hibit uniformity with other instances. Intelligibil- 
ity does not require such "through and through 55 
uniformity but only some uniformity. What such 
uniformity shall be, we do not dictate to experi- 
ence, save that specific kinds are requisite to sub- 
sumption in particular categories. Beyond that, 
the rational demand for uniformity wears some- 
what the appearance of regulative ideal, or maxim 
for the conduct of investigation. Intelligibility 
and understanding are not incompatible with irre- 
ducible variety; such lack of uniformity may be 
irrelevant to the particular mode of cognition. 
Further, we seem to have a theoretical interest in 
unlimited variety, as is evidenced by our require- 
ment of uniqueness in the individual thing. Again, 
much of the basic uniformity of various areas of 
experience is not discovered but imposed by cate- 
gorial procedures which argue nothing intrin- 
sically orderly in what is given. Outstanding ex- 
amples are the serial and dimensional orderings of 


qualitative variety, as well as schematic classi- 
fications, on the basis of similarity, by which an 
indefinitely large range of possible variety is 
brought within a definitely limited number of 

If now, I may suppose that this inadequate 
examination will be sufficient to guard against the 
commoner misapprehensions about the "uniform- 
ity of nature," I should like to proceed to those 
considerations which I believe to be really cen- 
tral for our problem. What is required in the way 
of order if experience is to be intelligible and 
knowledge possible is only that there should be 
apprehensible things and objective facts and to 
this we can conceive no alternative whatever, un- 
less it be the non-existence of everything. 

As should now be evident, the existence of an 
apprehensible thing is not assured by the mere 
givenness of experience, though if there were no 
things for us, there would be, in an obvious sense, 
no experience. Things exist for our apprehension 
as certain sequences of possible experience, of 
which given presentations are probable indices. 
For this, the sole necessity is that, certain presen- 
tations being given, the possibilities of further 
experience should not be unlimited; that is, that 
it should not be the case that every recognizable 
appearance is equally associated with, or followed 
by, every other. Let us give this fundamental re- 
quirement of knowledge formal statement: Prm- 


ciple A; It must be false, that every identifiable 
entity in experience is equally associated with 
every other. This principle is what Mr. Keynes 
has called the "limitation of independent va- 
riety,"* except that it is here applied to the iden- 
tifiable constituents in experience, particularly 
with reference to their sequence, instead of to the 
qualities of objects with reference to their corre- 
lation in reality. 

What I wish to point out is, first, that this sin- 
gle requirement satisfies everything which is nec- 
essary, in the way of order in experience or real- 
ity, for the validity of empirical generalizations, 
based on past experience and applicable to the 
future; and second, that although this has the 
appearance of a limitation of the possibilities of 
experience it has in the end no alternative* To 
put it in paradox; every possible experience is 
ipso facto a possibility of experience, but it is not 
possible that all possibilities should be actual. Any 
possibility is a possible actuality, but it is not pos- 
sible that all possibilities should be concomitantly 
real. The coincident actuality of all possibilities 
is impossible* Thus the requirement that actuality 
be a limitation of the all-possible, is not itself a 
limitation of the possibilities. Instead, it is some- 
thing which could not conceivably fail to be the 
case ; there is no alternative. 

Let us turn first to the requirement that there 

*A Treatise on Probability. Ch. XXH. 


should be apprehensible things and objective facts. 
As has been pointed out, the existence of things 
and the possibility of our knowing them in the 
only sense in which we can know them does not 
require that what we attempt to regard as things 
should be in each and every case objective reali- 
ties. Given appearance is no more than a probable 
index of the existence of such a particular reality. 
Furthermore, any particular substantive concept, 
even though it should be current for a thousand 
years, may eventually turn out to represent, not 
an objective thing, but a mistaken reification. 
The concatenation in experience which the ap- 
plicability of the concept requires may not, in 
fact, obtain- There may be no unicorns ; there may 
be no disease entities ; there may be no such thing 
as the soul. If the reality of knowledge required 
that every uniformity which we seek, by the in- 
vention of a concept to designate it, should be 
present in possible experience, or if it required 
that every given appearance should be an index 
to some uniformity which could be predicated 
with certainty, then it would not be plausible that 
there is any such thing as knowledge. All that is 
required is that given appearances should be 
probable indices to such uniformities as may be 
designated by concepts. This in turn means that 
statistical generalizations from the past sequences 
of experience must establish in some measure their 
probability for the future. And this, again, de- 


pends, as has been stated, upon the limitation of 
independent variety in the correlation between 
given appearances and further possibilities of ex- 

The point may be made clear by an analogy 
which, as the reader should be warned, is incom- 
plete but may nevertheless be useful. Suppose that 
we observe a certain sequence in cards, dealt from 
a pack, to be repeated several times. Will this es- 
tablish a probability of its future repetition? If 
we may suppose some limitation of the possibili- 
ties, such as adhering of the cards or trickery in 
the deal, it may. But if we suppose no such limi- 
tation upon those ideal conditions which packs of 
cards and their shuffling are meant to approxi- 
mate, then even numerous repetitions of an ob- 
served run will not iix the least increase the ante- 
cedent probability of its future occurrence. If, 
then, we may compare identifiable qualia and com- 
plexes of such (recognizable appearances) to the 
cards, and experience in general to an ideal pack 
under ideal conditions, a statistical generaliza- 
tion of past sequences in experience could not es- 
tablish that in future one such sequence should be 
probable as against others in, which the given ap- 
pearance should be the same. That is, no appear- 
ance would be probable index to further possible 
experience. Since what any substantive concept 
denotes in experience is some definite sequence or 
set of such, there could be, in an experience which 


was subject to no limitation upon the all-possible 
sequence of given presentations, no recognition of 
things by their appearances as even probable. 

But to continue our analogy if we can sup- 
pose the pack from which cards are dealt to be 
defective or the dealing subject to trickery, then 
a statistical generalization as to past runs will es- 
tablish a genuine probability for the future, and 
the continued verification of such a generalization 
will continually increase its probability. It will 
still be the case that any particular generalization 
of this sort may represent "mere coincidence" in 
the past and fail in future. But the point which 
it is especially important to observe is that if the 
assumption that there are limitations which some 
uniformity of sequence might reflect is a war- 
ranted assumption, then this fact of itself is suffi- 
cient to establish the validity of the argument, in 
any particular case, from past uniformity to its 
probability for the future. If, however, it should 
be positively known that the runs of cards are 
ideally governed and subject to no such limita- 
tion, then there would be no such probability. The 
especial point here is that the validity of arguing 
from any past coincidence to its probability in the 
future does not depend upon our knowing any 
particular ground for this particular uniformity 
(other than its past occurrence) , but does depend 
upon our knowing as at least possible a ground 
for such uniformities m general. 


So far the point of our analogy is, that the 
validity of arguing from the correlation between 
a given presentation and that further experience 
which means a particular kind of object denoted 
by a particular concept, does not require us to 
know any particular reason for that particular 
correlation or for the existence of just that sort 
of thing (other than the past experience itself). 
It requires us to know as possible at least that 
such correlations, things m general, exist. If this 
assumption is valid, then the prediction (as prob- 
able) of that future experience which means the 
verification of the presence of an object corre- 
sponding to a particular concept, from a given 
presentation with which such further experience 
has been associated in the past, is valid ; and the 
probability is equally genuine when, as a fact, the 
past conjunction is a "mere coincidence" which, 
future experience will prove not to be valid as a 
generalization. That is, mistaken apprehensions, 
as probable judgments^ will, if there has been no 
logical error in the judgment, be valid, as much 
as those which are actually verified by the future. 
It is also to be noted that, if the assumption 
that things in general exist should be valid, then 
the probability for the future of the correlation 
between a given presentation and certain further 
experience will be increased with each successive 
verification of that correlation. The particular 
principles of probability upon which this rests 


need not be examined here;* the reader will per- 
haps be satisfied upon the point by consideration 
of the analogy: if we have ground for assuming 
poor shuffling or interference or some such limi- 
tation, the probability for the future of a par- 
ticular run which has been repeated will increase 
with each repetition of it. At least this will be so 
after a certain number of repetitions have al- 
ready occurred, if we suppose the conditions un- 

This assumption of the existence of things, that 
is, of certain recurrent correlations in the se- 
quence of possible experience, is all that is re- 
quired for the validity, as probable, of empirical 
generalizations or "laws, 5 * and of the argument 
from past to future with respect to these. It will 
be evident from previous chapters that the ex- 
istence of objective properties is conditional upon 
precisely this same type of uniform sequence in 
experience. The verification of such properties re- 
quires precisely the same connection between given 
appearances and further possible experience, and 
requires nothing more. Obviously the assumption 
of objective properties is involved in the assump- 
tion of things. Furthermore, if things exist, which 
are cognizable and therefore the objects of pos- 
sible concepts, it is evident that there must be 
laws. Laws prescribe, or describe, precisely such 
uniform sequences. In fact, although laws or em- 

*On this point, see Keynes, "A Treatise on Probability." Ch. XX. 


pirieal generalizations formulate relations of ob- 
jects, or properties of objects, which are non-es- 
sential rather than essential, this difference is 
irrelevant to the possibility of knowledge. What 
is essential in a thing is determined by the par- 
ticular concept which it serves our interest to 
apply, rather than by anything else. The proper- 
ties essential to "a stone" do not include its being 
"a freely-falling body," and the essential proper- 
ties of a "freely-falling body" do not include its 
being a stone. But certain laws or generalizations 
must hold of an object in order that it be a stone 
and certain others must hold in order that it be 
a freely-falling body. The laws which character- 
ize or constitute essential properties of freely-fall- 
ing bodies are non-essential or merely empirical 
generalizations about stones under certain cir- 
cumstances. Moreover, not only are all essential 
properties capable of representation as laws, but 
every empirical generalization is such that some 
substantive concept is capable of being framed so 
as to require conformity to it as the distinguish- 
ing or essential property of a particular kind of 
thing. Scientific concepts especially have this 
character explicitly; they define or classify ob- 
jects on the basis of their conformity to certain 
laws, prescribing modes of behavior. Indeed, any 
objective fact meaning by this any state of af- 
fairs which is describable in conceptual terms is 
a property of objects and may be denoted by a 


substantive concept marking it as essential. In 
fact, unless the whole point of the last two chap- 
ters has been lost, it will be clear that the differ- 
ence between essential and non-essential proper- 
ties, and between prescriptive laws and empirical 
generalizations, is one determined by pragmatic 
considerations of the particular interests our knowl- 
edge is to serve which dictate particular modes of 
analysis, rather than by any difference in the ob- 
jective state of affairs upon which these are di- 
rected. What laws must be valid, would depend 
upon what things exist; but the general assump- 
tion that there are things (of some sort) includes 
the assumption that there are valid generaliza- 
tions of the type of law. If there could not be a 
world without the uniformities of possible experi- 
ence requisite to the existence of things, then there 
could not be a world without uniformities of the 
type of law. 

As is the case with things, so too with law, the 
validity of the prediction that a particular gen- 
eralization which has held in the past will likewise 
hold in future, depends upon the general assump- 
tion that there are laws. If there are such uniform 
sequences as laws describe, then the occurrence of 
one such in the past establishes a probability of 
it in future, and each successive verification in- 
creases that probability. And here, too, the prob- 
ability-judgment, if logically drawn, Tpill be valid 
even in those particular cases in which future ex- 


perience will prove past instances to have been 
"mere coincidences' 5 and the generalization to be 

Thus the assumption that things exist that 
there are (some) such recurrent sequences as sub- 
stantive concepts require for their application 
is sufficient to secure the validity of knowledge in 
general, when the nature of empirical knowledge 
is correctly interpreted as probable judgment. As 
has been remarked previously, the form which 
skepticism has often taken, of doubting the valid- 
ity of all empirical generalizations while leaving 
unquestioned the existence of apprehensible things, 
in terms of which such dubious generalizations 
could at least be intelligibly phrased, represents a 
totally impossible position. It could only seem self- 
consistent on the assumption that the validity of 
knowledge requires the certainty of empirical gen- 
eralization and of prediction. Since any reason- 
able examination of knowledge must conclude that 
the pretense to such certainty is unwarranted and 
the ascription of it is a misreading of the actual 
nature of science and of common-sense attitudes, 
what such skepticism has slain is a man of straw, 
though to be sure it is just this scarecrow which 
has frightened philosophers out of their wits for 
a considerable period. 

Before passing on, we should remind ourselves 
once more at this point that the assumption of the 
existence of apprehensible things does not mean 


the assumption that experience is uniform in the 
sense that certain determined sequences univer- 
sally follow upon a given first term (the presenta- 
tion). If it meant that, then no apprehension of 
a thing would be valid unless the given presenta- 
tion should be such as to render mistaken ap- 
prehension absolutely impossible. Since errors of 
apprehension are possible and identification of ob- 
jects are probable, it follows that the prescribed 
sequences, requisite to the existence of things, are 
such as would be predictable, with certainty, only 
in terms of some larger whole of experience than 
can be included in a single presentation. Our ac- 
tual predictions our actual knowledge of things, 
predicated upon given presentations is in terms 
of such sequences of experience as admit of ex- 
ceptions such as mistake and illusion. Hence as 
generalizations about actual experience, the ex- 
istence of apprehensible things means only such 
uniformity as can be formulated in generaliza- 
tions of the statistical type, which admit of excep- 
tion. No absolute wniformities of actual experi- 
ence are required either for the existence of things 
or for the objective character of law. Laws too, 
of course, are exempt from being proved false by 
mistaken apprehensions of the things or objective 
facts with which we suppose ourselves to be deal- 

The examination of the categories, "thing," 
"event," "property," "relation," "law," is a most 


difficult and complex matter. I make no pretense 
to have exhausted even the essential considerations 
for any one of these. In particular, I should not 
like to seem to assert that there is no difference 
between what is requisite to the nature of a thing 
and to the objectivity of law. The attempt is only 
to show that there could not be a world of appre- 
hensible things in which empirical generalizations 
should fail of valid foundation; that if there are 
things then laws of the type which empirical gen- 
eralization seeks to grasp must hold; and hence 
that such generalizations may be genuinely prob- 
able and that empirical knowledge, as the only 
sort of thing it can reasonably be supposed to be, 
is genuinely possible. 

If this point is established, then the only al- 
ternative to the conception that our knowledge in 
general is valid, is the conception that there are 
no things; that nothing exists to be known and 
no mind exists to know it. The nearest approxima- 
tion we can make to such a conception is, perhaps, 
that there might be an experience which is a mere 
flitting of meaningless presentations. But for such 
experience, if we can conceive it, the distinction 
of "real" and "unreal" could have no meaning. 
This being so, it is a little obscure just what we 
suppose ourselves to be talking about when we try 
to frame such a conception; perhaps about the 
experience of an oyster with the oyster left out. 

I do not mean to take advantage of the fact 


that skepticism can only make itself intelligible in 
terms which render it not self -consistent. This is 
a fact, and a most important one. But it might be 
claimed though I do not know that it ever has 
been that skepticism intends only to exhibit a 
reductico ad dbsurdum of the pretense to knowl- 
edge, by beginning with definite assumptions 
about the mind, etc., and on that basis proving the 
invalidity of these assumptions along with every 

This possible intent of skepticism seems the 
more important to examine because it would ap- 
pear, in the above, that the account of knowledge 
here given avoids skepticism only by an assump- 
tion which has a conceivable alternative ; and that 
this alternative is precisely the absence from ex- 
perience of all order of the kind which means sig- 
nificance and intelligibility. In this situation, it 
might be said: "It is not humanly possible to di- 
vest ourselves of assumptions such as the exist- 
ence of definitely conceivable things which have 
no rational foundation, and still talk or think, but 
it is possible to be sufficiently self -critical to real- 
ize that such unavoidable assumptions are non- 
rational and mere 'animal f aith.* " 

I believe this to be demonstrably false, and shall 
attempt to make this fact clear. But before ap- 
proaching that topic, there are two points about 
skepticism, with the above meaning, which it may 
be well to observe. 


Historically there have been two main grounds 
of skeptical conceptions, the relativity of sense- 
perception and the absence of "necessary connec- 
tions" in experience. Quite often these are con- 
fused together and results appropriate only to the 
first are added to the second. It is the first of these 
only which gives plausibility to the conception 
that there is a reality which is unknowable to us 
because we are separated from it by the manner 
of our apprehension. This ground of skepticism 
has, I hope, been dispelled by the considerations 
of previous chapters. It neglects the fact that 
"real" is systematically ambiguous ; that "appear- 
ances" themselves must constitute one kind of re- 
ality. It also neglects the further facts that reality 
of any sort is definable and meaningful only in 
terms of some experience, actual or hypothetical, 
and that regardless of the relativity of percep- 
tion, appearances inevitably are, for a rational 
understanding, a ground of true knowledge of the 
reality which appears even though that knowledge 
should be incomplete. As a result, it is impossible 
to conceive "reality" as completely unknowable; 
and since it is not plausible that, under actual or 
realizable conditions, reality can be completely 
known to us, the significance of the "unknowable" 
dwindles to the commonplace fact that humans 
are not omniscient. Nothing in the train of thought 
which starts from the relativity of perception can 
in any way vitiate such knowledge as we have or 


seem to have. The ground of this sort of skepti- 
cism is the false conception of knowledge as rep- 

The other type of skepticism turns upon a cor- 
rect conception of knowledge as predictive judg- 
ment. Its particular ground is the absence of "nec- 
essary connections of matters of fact." This may 
well be equivalent to the falsity of the assumption 
which has here been shown requisite to the exist- 
ence of apprehensible things. Actually, of course, 
Hume supposed that necessary connections must 
mean such iron-clad uniformities in experience as 
would enable certainty of prediction, whereas it 
is only genuine probability which is requisite. But 
I hope it will be clear that if this is the ground of 
skepticism, it is quite unwarranted to supplement 
this conception by a notion of a reality somehow 
concealed from us by the chaotic character of ex- 
perience. The logical conclusion would be that of 
Gorgias, "Nothing is," or at least the admission 
that we can have no rational ground for asserting 
a reality of any sort. If we cannot humanly avoid 
this, that is merely to observe the fundamentally 
irrational character of our "animal faith" and the 
reductio ad dbsurdum of all attempts at consistent 

The second point to be observed before we pass 
on, is that we must not confuse such irrational 
"animal faith" with that wholly rational attitude 
of him who, acting on the basis of a probable 


judgment, confronts the future he predicts with 
the realization that the possibility of his heing 
disappointed is a real one. In both cases, one would 
face the future with an attitude determined by 
past experience but one which, in the particular 
instance, the future might prove to be fruitless. 
But in the one case, this attitude has no rational 
ground and incorporates no truth; in the other, 
it represents a knowledge of probability which 
is not only rational but absolutely true, whatever 
the denouement. If one should ask, "But what, 
practically, would be the difference?" that is a 
point which we shall reach shortly* 

We come now to the main point: it is impossi- 
ble to imagine any sort of experience which would 
not present such statistical stabilities as would 
validate probable prediction, and such as would 
represent the experience of things- Our analogy 
of experience to runs of cards is, on this point, as 
unfavorable as could easily be devised, because 
packs of cards and their manipulation are intelli- 
gently directed to the end of minimizing the pos- 
sibilities of prediction in those respects which de- 
termine the outcome of the game. But it is worth 
remarking at the outset that the point of card- 
games, except for children, is to pit our wits 
against this maximal uncertainty and determine 
a long-run outcome favorable to ourselves in spite 
of it. If the sequence in experience should be as 
independent and lacking in "necessary connec- 


tion" as the sequence of cards, that could not frus- 
trate prediction or destroy the practical value of 
probability- judgments. Indeed if cards should 
represent, in the analogy, recognizable qualia or 
complexes of such in presentations, our attention 
should be drawn to the fact that, on this point, 
the analogy is better than might be supposed. 
There are no sequences in experience which are 
determinable with certainty by a given presenta- 
tion alone. Real things represent stabilities of a 
type which enormously transcend what can be 
given in any one experience. If the reality of 
things required the presence in experience of se- 
quences absolutely determined by a first term, then 
we should have no reason to believe in their ex- 
istence. As has been pointed out too often already, 
what is required is only that a given presentation 
determine a probability of future possible experi- 

The reason why Principle A imposes no actual 
limitation upon the possibilities of experience may 
be formulated in a second dictum: Principle B; In 
any situation (if sufficiently extended) in which 
there are identifiable entities which fail to satisfy 
Principle A i. e., whose association is "random" 
there will be other entities, systematically con- 
nected with the former or specifiable in terms of 
them, which do satisfy Principle A. The principal 
methods by which we determine "orderly" con- 
stituents of experience in terms of "random" ones 


are by proceeding to simpler elements through 
analysis, by taking a larger whole into which the 
primary constituents may be organized, or by con- 
fining attention to abstracted elements, and dis- 
regarding the remainder of the given as irrelevant. 

To revert to our analogy, if the sequences of 
cards are purely random or subject to no laws but 
those of chance, then the stable entities of card- 
games, comparable to things, will be something 
such as tricks-taken, or suits, or kinds of hands 
such as full-houses, straights, etc. Or they will be 
entities of a lower order such as the pips on the 
cards. As is obvious, the chance character in the 
sequence of dealt cards does not frustrate the at- 
tempt at statistical generalizations which give 
guidance for successful play. Even with entities 
thus deliberately devised to approximate pure 
chance in certain ways, there are certain wholes 
which, neglecting the "non-esential," give rise to 
generalizations. In fact, no better basis for sta- 
tistical generalization could be devised than the 
known fact that certain constituents of the situa- 
tion are distributed in genuinely "random 35 fash- 
ion* Since any departure from "pure chance" is 
itself subject to generalization of some sort, a sta- 
tistical basis for probable judgment cannot con- 
ceivably fail to be afforded.* 

What particular stabilities, and what types, are 

*In this connection, the analogy will be improved if we think of our 
predictions about runs of cards as based on past runs (as they might 
be), not upon prior knowledge of the constitution of the pack. 


to be found, cannot be prescribed to experience or 
reality. The particular order discoverable in re- 
ality, the extent of it, and the degree of its con- 
ceptual simplicity, are of course, absolute data. 
The supposition that we may always find law in 
terms of any experimental entities we please and 
with predetermination of the type of the uni- 
formity to be found, is totally unwarranted and is 
unnecessary to the validity of knowledge. Our con- 
cepts are devised with purpose to catch the sig- 
nificant, the subject of meaningful generalization, 
at whatever level and in whatever way we may. 
When particular concepts fail, we merely abandon 
them through analysis or organization or ab- 
straction, and so on in favor of corrected ones, 
which take cognizance of, and include the ground 
of, our previous failure. That conception in gen- 
eral should be invalid, is quite impossible. The at- 
tempt to envisage an experience or state of affairs 
such that every attempt to discover stabilities must 
fail, is the attempt to conceive the inconceivable 
to conceive what would not be things or objec- 
tive facts nor subject to any generalization which 
makes what is denoted conformable to concepts. 
The experience or reality which should be incom- 
patible with conception, ipso facto cannot be con- 

It may seem that there is one aspect of the mat- 
ter not yet covered ; the validity of the argument 
from past to future. It may appear that it would 


be possible, not only that particular predictions 
based upon the past should be useless as anticipa- 
tion of the future which their character as prob- 
able allows but that in general the anticipation 
of the future in the light of the past is without 
theoretical warrant. This point really is covered 
by Principle A, when we remember that the kind 
of association of constituents in experience which 
is essential to their comprehension in things is se- 
quence in possible experience. But when we re- 
member that it is the validity of probable predic- 
tion only which is required, the matter can be 
more explicitly stated, and quite simply : Principle 
C; the statistical prediction of the future from the 
past cannot be generally invalid, because what- 
ever is future to any given past, is in turn past 
for some future. That is, whoever continually re- 
vises his judgment of the probability of a statisti- 
cal generalization by its successively observed veri- 
fications and failures, cannot fail to make more 
successful predictions than if he should disregard 
the past in his anticipations of the future. This 
might be called the "Principle of statistical ac- 
cumulation." It is quite evident that it holds even 
with respect to what is determined only by "pure 
chance" in the only sense that we can conceive 
anything such. This is what is meant by saying 
that probability or chance is measured by that 
fraction which is approximated "in the long run." 
Though the attempt to envisage what would 


frustrate conception and knowledge is, in the na- 
ture of the case, futile, it may be worth while to 
conceive the worst possible experience, from that 
point of view, by a somewhat fantastic illustration. 
Let us think of our experience as constituted by 
complexes or patterns of qualia which come to us 
in sequence, and let us suppose that this experi- 
ence is given to us, not by a Berkeleian God who 
in his goodness preserves certain uniform sequences 
in order that we may predict experience accord- 
ing to natural law, but by a perverse demon whose 
sole purpose is to mislead us and render knowl- 
edge impossible* If the distinguishable sensory 
qualia should be finite in number, as they are in 
actual experience, this demon must necessarily re- 
peat, but he might repeat identifiable complexes 
and previous sequences to as small a degree as pos- 
sible. However, if, as a result, these presented pat- 
terns did not afford a sufficiently good basis for 
conception, we should analyze them into sub-pat- 
terns or other constituents which would be of more 
frequent appearance, or should classify them ac- 
cording to some simplified schematism based on 
similarity, or give them dimensional or serial ar- 
rangement approximating to an ideal continuum, 
or in some other fashion proceed by abstraction or 
reorganization or a combination of the two to cir- 
cumvent the multifarious variety of the given. For 
the rest, the character of it would be relegated to 
the "non-essential" and merely mark the relative 


uniqueness of particular moments of experience; 
it would be non-significant, like the particular 
character of frost-patterns and snow crystals. 
We should be inattentive to such non-significant 
features of the given; and if any one should call 
our attention to their comparative frequency, we 
might regard his observation as a foolish remark 
about the obvious but unimportant. If the demon 
should likewise minimize the extent to which par- 
ticular sequences should be repeated, he might 
make knowledge difficult to a degree. But at least 
we should presumably come to possess the very 
important generalization that the maximum of 
novelty may confidently be expected. We should 
organize all our conduct on the principle that 
"lightning seldom strikes twice in the same place" 
and "history never repeats," with consequent ad- 
vantage to ourselves. As a fact, however, we should 
circumvent the demon on this point at the same 
time that we attained relative simplicity of recog- 
nition, and should carry out similar procedures 
with reference to sequence. We should analyze, 
abstract, and relegate what we could not find 
somehow significant to the status of "irrelevant 
concomitants of significant 'causal* sequences." 
By ignoring a sufficient proportion of the charac- 
teristics of experience as it came to us, we should 
arrive at such simplicity that, in terms of it, even 
the most disadvantageous sequence of the primary 
constituents e. gr., a "random" order must af- 


ford some repetition and uniforgtiity. Knowledge 
might be made difficult, but could not be made im- 

This would be something as if we were required 
to play a game with cards dealt from a pack we 
never saw and could only infer from the hands 
dealt us, and as if these cards were dealt with 
trickery and malice something worse than a 
good decent game of chance. But even in such 
circumstances, the principle of statistical accu- 
mulation would operate. If we were required to 
bet on this game, we might be unequal to our 
demon antagonist to any degree you please; our 
ignorance might be great, and our failures and 
loss correspondingly large. But by nothing which 
he could do could be so devise it that we should 
not lose less of our money if we intelligently ob- 
served past dealings and continually revised our 
betting on the basis of accumulated experience. 

Indeed we need only to prod our imaginations 
to remark that this actual experience of ours fits 
the illustration better than one might suppose* 
Most of the pattern and sequence of our experi- 
ence is non-significant ; those characters which by 
our practical attitude we single out and remark 
and make prediction in terms of, are inordinately 
meager as compared with the total of presented 
distinguishable pattern and sequence of qualia. 
Most of experience is as non-significant as the in- 
tricate pattern of the rug I have been gazing at 


is to me now. Beyond a few details which, suffice 
for recognition, this intricacy does not even draw 
attention, and if attended, it hardly functions as 
knowledge. It neither verifies anything nor baffles 
knowledge, because it arouses no anticipatory at- 
titude. Confronting any given experience, the first 
act of intelligent cognition is to discard all but a 
few items of what is presented as excess mental 
baggage irrelevant from the point of view of our 
predictive purpose. It is the relatively meager re- 
mainder which constitutes the clue to expected 

Even if the hypothetical demon should have an 
infinite number of qualia and complexities at his 
disposal, instead of a finite number, it is doubtful 
if we should be worse off. We should merely dis- 
card more of our experience, as marking that 
uniqueness of each moment and each thing which 
is irrelevant to knowledge; we should frame our 
concepts and make our predictions in terms of the 
remainder, or of abstractions and other such im- 
posed simplifications. On any hypothesis which 
can be framed about experience, however per- 
verse, I think it will be clear that a similar thing 
would hold, or else that the fantasy is not even im- 
aginable but would merely mean the elimination 
of reality, of significance, of all questions and the 
mind itself. 

In any experience such as we can, even at the 
worst, suppose our own to be, conception will be 


valid and knowledge will be possible. The three 
principles which have been stated will hold, and 
generalization will be subject to genuine proba- 
bility. No further and avoidable metaphysical as- 
sumption is required* The mind will always be 
capable of discovering that order which is requisite 
to knowledge, because a mind such as ours, set 
down in any chaos that can be conjured up, would 
proceed to elicit significance by abstraction, analy- 
sis and organization, to introduce order by con- 
ceptual classification and categorial delimitation 
of the real, and would, through learning from 
accumulated experience, anticipate the future in 
ways which increasingly satisfy its practical in- 



In his introductory chapter to "Einstein's The- 
ory of Relativity/ 5 Max Born has written:* "The 
development of the exact sciences leads along a path 
to a goal which, even if far from being attained, yet 
lies clearly exposed before us : it is that of creating 
a picture of nature which, confined within no limits 
of possible perception or intuition, represents a pure 
structure of conception, conceived for the purpose 
of depicting the sum of all experiences uniformly 
and without inconsistencies/* And he goes on to 
characterize this world of "inaudible tones, invisible 
light, imperceptible heat, 5 * in which the limited range 
of the human senses is ignored, as a "sum of ab- 
stractions** and "subtle logical configurations.** 

It may well be that the inevitable movement of 
the exact sciences is in this direction in which mathe- 
matics has preceded them; toward the deductive 
mode of development and toward concepts which 
are laid down less in terms of those sense-qualities 
by which we directly identify empirical objects and 
more in terms of those systematic correlations 
which figure in natural law. The physical definition 
of sound or color in terms of wave-lengths or rates 
of vibration, serve as simple illustrations of the sort 
which Born has in mind. Concepts which are ab- 
stract to a degree, and remote from the merely ap- 
parent or directly discriminable, are dictated by 
interest in that uniformity and comprehensiveness 
which are essential to intellectual economy and ex- 

*English. translation, p. 2. 



actness. And the deductive method of presentation 
as we are beginning to understand is not a 
method of proof at all (since the "first principles' 5 
are as much corroborated by their consequences as 
the consequences are by their deducibility) but is 
simply the most compendious and economical method 
of tying up an enormous multiplicity of facts in a 
relatively small number of bundles* It is a method 
which becomes possible only when a high degree of 
precision has been attained and the interrelations 
of different classes of phenomena are quite thor- 
oughly understood. But when these requirements are 
met, the power of deductive order to summarize 
great masses of facts, while still preserving their 
relations to one another, is an advantage almost 

Abstractness in the concepts, and that systematic 
order which reaches its highest degree in the deduc- 
tive system, go hand in hand. The reason is one 
which is inherent in the nature of the problem which 
science undertakes to solve* The conceived "things'* 
of the unsophisticated consciousness are, relatively 
speaking, coagulations of sense-qualities; they are 
such as are identifiable, with a minimum of risk, by 
their momentary appearance. But thinghood must 
also include objective change, since the purpose of 
such abstractions, or excisions, from the immediately 
presented is the possibility of prediction. Science 
but seeks to raise to a higher power this possibility 
of correlating identifiable "thing" and predictable 
change. In so doing, however, it is obliged to aban- 
don, in some measure, the things of common sense, 
relatively identifiable by their appearances, and to 
substitute therefor things which are conceived in 
terms of correlations less directly observable. In its 
basic and defining concepts, it moves away from 
sense-appearance and in the direction of law. Those 
systematic correlations, for example, which are first 
painfully established as the uniform behavior of 


gravitating objects a law of change later become 
the defining characteristics of "mass," which in turn 
becomes the essential property of matter. Or those 
uniformities which chemistry pursues are finally at- 
tained in terms of atoms or electrons (which are es- 
sentially imperceptible, and in that sense purely con- 
ceptual entities). Such ultimate things, identifiable 
by sense only through uniformity of behavior and 
correlation, displace the tastes and odors and other 
sensible qualities by which salts and acids, gold and 
air and water, are identified. And then, perhaps, we 
mystify ourselves because we have a world of "in- 
audible tones, invisible light," and matter that we 
cannot imagine! 

This regress of science from the directly percep- 
tible may be phrased in another way. Any "thing" 
is a bridge between given experience and predictable 
change* A thing of any sort must always be such 
that it is identifiable directly or indirectly, with 
easy assurance or with difficulty by something 
which is the "appearance" of it. Else the concept 
would meet with no clue to its application, and the 
purpose of it would be lost. But thinghood must also 
include objective change; otherwise the identifica- 
tion of it in direct experience would enable no pre- 
diction and, once again, the purpose of the concept 
would be lost. The difference between the concepts 
of unsophisticated common sense and those of sci- 
ence is that common-sense things are relatively easy 
to identify but relatively unreliable guides to pre- 
diction, whereas the things of science are relatively 
safe guides to prediction but correspondingly diffi- 
cult of immediate identification. This is because the 
scientific concept takes that correlation and uni- 
formity which may be formulated as law and makes 
it the essence of the thing. 

I should not like to become entangled with the 
ancient and honorable notion of "substance," but it 
is evident that we may find here a clue to some of 


the problems concerning this. Since a concept of 
that which is merely momentary would serve no pur- 
pose, the "thing itself" (for any grade of thing) is 
always conceived by means of those properties which 
persist through the process of experience or objec- 
tive change. These are its essential properties. When- 
ever this relatively stable complex of properties is 
altered, the thing in question goes out of existence 
e. g. 9 the wood is burned, leaving ashes. But the 
extreme antithesis of being a "thing" is not being 
an objective change which is subject to law since 
what is predictable may be brought within the con- 
cept and made essential but is merely such change 
as should be unpredictable, uncontrollable, and 
hence baffling to the understanding. Thus when the 
change is predictable or controllable, it is not a 
"disappearance" but a "transformation," internal 
to the nature of some conceivable "thing." The 
predictable transformation from liquid to solid at 
32 F. is an essential property of water. It is not a 
property of that potable liquid which precedes the 
freezing, and is annihilated and replaced by ice. (At 
least it is not such for an unsophisticated tropic- 
dweller, whose concept cannot include prediction of 
it.) It is a property of that thing whose concept is 
satisfied both before and after the freezing; hence 
of that which persists and is the subject of the trans- 

Thus in every transformation that is, in every 
process of objective change which is predictable and 
hence intelligible there are two layers of things. 
The lumber disappears, the desk comes into being; 
but it is the same wood. Some of the essential prop- 
erties of lumber have disappeared, some of the es- 
sential properties of a desk have come into being; 
but all those of wood have persisted. When our last 
word is "It disappeared" or "It came into being," 
we cannot predict or control change, which is what 
we mean when we say that we cannot understand it. 


Destruction and creation are, thus, unscientific cate- 
gories. The methodological postulate of all science 
is that the problem of understanding the process of 
reality, to which it addresses itself, is essentially ca- 
pable of solution; and hence, that all change is 
transformation. Thus the ultimate things of science 
must, in the nature of its ideal, be eternal; and it 
cannot stop until it has disclosed them. That in its 
attempt to conceive what is eternally persistent, it 
must import into its concept as essential the law 
which anticipates transformation in general, and 
that the ideatum of such a concept will be remote 
from what can be directly identified in sense-experi- 
ence, and perhaps will not be imaginable at all, be- 
longs to the nature of the case. 

Nevertheless it remains true that the difference 
between the substantive concepts of science and 
those of common sense can only be one of degree. 
As has been pointed out, all conception of things is 
predictive; and on the other hand no concept of 
science would be significant if the thing denoted 
could not be indirectly identified in experience 
through its manifestations, correlations, or effects. 
The point to be noted is only that the relative ab- 
stractness of scientific concepts is an inevitable 
concomitant of its greater comprehensiveness and 
power of accurate prediction. Increasing abstract- 
ness and an increased satisfaction of the ends pro- 
jected by "exactness," "order," "system," go to- 
gether. Both are consequences of that formulation 
of comprehensive key-conceptions in which basic 
things and fundamental laws are no longer distinct 
ideas, because the substantive concepts are them- 
selves framed in terms of those correlations of phe- 
nomena exhibited in law. 

In the march toward these ideals, there comes a 
stage when it is no longer easily possible to say 
whether concepts are devised, and laws discovered, 
to fit phenomenal facts, or whether the conceptual 


system itself rules and facts are reconceived in con- 
formity to it. At least the give and take between 
these two is on approximately equal terms. Whether, 
for example, the devising of such concepts as "en- 
ergy of position" and "curvature of space" repre- 
sents a modification of system or law to fit facts, or 
an alteration in the manner of conceiving facts so 
as to fit an a priori comprehensive schema of inter- 
terpretation, seems to be a question only of that 
aspect of the procedure which shall be emphasized. 

With exact natural science, as with mathematics, 
a stage is possible, if it is not already reached, in 
which the problem of scientific truth can be phrased 
equally well as the discovery of empirical laws suffi- 
ciently comprehensive to constitute a systematic 
whole, or as the selection of an abstract system 
which will be applicable to the facts. Just as the 
problem of space may be envisaged either as that of 
discovering the system of laws governing certain re- 
lations of ideally rigid bodies, or as that of selecting 
an abstract geometry which accords with observed 
phenomena, so perhaps the issues between a New- 
tonian and a relativity kinematics may be phrased 
equally well as questions of the correct generaliza- 
tion of certain physical phenomena or as the ques- 
tion of choice in application between the two ab- 
stract systems, both of which have that logical in- 
tegrity which comes from strict adherence to their 
fundamental concepts. It would be but a short step, 
if it is not already possible, to viewing such alterna- 
tive systems as strictly deductive elaborations of 
purely abstract concepts and postulated relations. 
So viewed, such systems would have the same kind 
of truth, and be objects of the same kind of knowl- 
edge, which characterize a pure geometry apart 
from its applications. 

But if we thus extrapolate along the line of de- 
velopment which exact science seems to follow, and 
assign to it the highest degree of independence of 


directly given data of sense, it still remains true that 
the truth about nature cannot have such indepen- 
dence. When we inquire upon what ground the se- 
lection of an abstract system to be applied to con- 
crete physical phenomena would be determined, it 
becomes clear that, directly or indirectly, sense-data 
must necessarily figure in such a decision. The logi- 
cal integrity of an abstract system is no guarantee 
of applicability. Let the connection between what is 
presented in sense and the idealized abstractions of 
the system be as remote as you please, this connec- 
tion is of the essence of any truth about phenomenal 
Scientific concepts import into themselves, and 
make essential to the scientific thing, more of what 
belongs to the systematic interconnections of phe- 
nomena ; and by way of compensation, they extrude, 
as non-essential, something of what is more apparent 
or easily observed. Now whatever is of the essence 
of a thing need not be established by induction or 
vested in empirical generalization. If, for example, 
possible resolution into hydrogen and oxygen is of 
the essense of water, there is no problem of induc- 
tion to establish this property. That something 
which, in ancient Greece, would have been classified 
under water,* as one of the four elements, might not 
be thus decomposable, has nothing to do with the 
matter, except as it marks the fact that we repudiate 
this ancient concept. 

If we reflect a little upon the history of chemis- 
try, or any other of the older sciences, will it not be 
obvious that the determination of what belongs to 
the deductive elaboration of concepts and what to 
empirical generalization from experience, depends 
rather simply upon our modes of conception them- 
selves? And further, that the manner in which sci- 
ence departs from common sense is characteris- 
tically one which enlarges the scope of deduction by 
the direction in which it modifies its concepts ? That 


such scientific concepts are built upon inductive 
generalization in earlier stages, is obvious enough. 
And that we cannot lift ourselves by our scientific 
bootstraps and enlarge our understanding of nature 
by altering definitions, will likewise need no com- 
ment. But it would be a misunderstanding to sup- 
pose that established principles of science begin as 
hypotheses or tentative generalizations and are, by 
continued inductive verification, finally made cer- 
tain, so that we dare embody them in definitions. 
More accurately, the process is one in which what 
we call induction enlarges the scope of experience, 
so that those correlations which are most useful for 
knitting together the facts of nature in a compre- 
hensive network are gradually revealed and con- 
firmed. Thus a more judicious ground of conception 
is reached. But the principles made definitive by 
such pragmatically superior conception neither wait 
upon any novel certainty to be established by induc- 
tion nor do they acquire such certainty when made 
definitive. As applicable to certain phenomena of 
nature, they are not completely certain either be- 
fore or after. When this potable liquid called water 
has been decomposed into hydrogen and oxygen a 
certain number of times without exception, the 
probability that this will always happen reaches 
such a pitch that this property may reasonably be 
adopted as the essential mark of a new concept. 
That another such experiment still might fail, is as 
possible after that as before. But it is now the pos- 
sibility that the (new) concept water will prove in- 
applicable to some tasteless liquid, instead of the 
possibility that water (as previously conceived) may 
fail to have this property universally. 

If we follow this process to the limit, we find that 
by gradual transition that stage of systematic cor- 
relation may be reached in which we have all the 
major principles of some science embraced in a de- 
ductive system. But this means no absolute cer- 


tainty about nature which did not previously exist. 
The problem now becomes whether, or how far, this 
deductive system is applicable to empirical facts. 


The nearest approach to pure givenness is doubt- 
less the esthetic experience. Cognition is always in 
part instrumental or pragmatic, but so far as an 
experience is esthetic, it constitutes an end in itself. 
Or if we use the word "esthetic" more widely, so as 
to include negative as well as positive values, at least 
the esthetic aspect of experience concerns its quality 
as end. 

There are any number of questions of esthetic 
theory which might be raised here: whether intrin- 
sic value is always a dimension of the immediate; 
whether esthetic "form** is given or is in part con- 
struction; whether the value-quality of the experi- 
ence should be regarded as intrinsic to the content 
or as a function of the mind or as a relation of the 
two. I beg leave to avoid such controversial ques- 
tions, so far as possible. It would seem that the 
facts with which we are concerned are fairly clear, 
and that such questions are, in part at least, such 
as will be settled by determining the field of esthetics 
rather than by answering questions about the na- 
ture of experience. It is obvious that the instrumen- 
tal sign-function of a presentation and its sig- 
nificance as immediate felt quale are cognate aspects 
of experience. These two cannot be separated in 
their temporal existence; every presentation has 
both at once. But the value of it as a cognitive sign 
and its value as esthetic are (presumably) inde- 
pendent. The former is extrinsic; the latter intrin- 
sic. Not only does all experience possess an intrinsic 
value-aspect a positive or negative value as an end 
in itself but "there is one glory of the sun and an- 



other glory of the moon": esthetic apprehension is 
not exhausted by placing the experience in a one- 
dimensional scale of immediate values. 

There is such a thing as direct appreciation of 
the given, and such immediate apprehension of the 
quality of what is presented must figure in all em- 
pirical cognition. Nevertheless the object of es- 
thetic judgment must always transcend the merely 
given. Given experience would need no appraisal, 
nor could the assessment of value exercise any help- 
ful function, if the object of it were simply and 
solely this experience itself. Experience wears its 
own intrinsic value-aspect on its face, and no ap- 
praisal of it as just this unique and given experi- 
ence is necessary. The object of appraisal is (usu- 
ally at least) to connect this quality with some 
thing or context as a matrix of further such expe- 
rience.* A judgment of value has direct relation to 
our action. That whose value is positive is to be 
sought ; that whose value is negative is to be avoided. 
But experiences having a certain quality can be 
sought or avoided only through just such prediction 
as is involved in empirical cognition in general. 

The most primitive of esthetic apprehensions are 
those expansive movements and approaches toward 
the stimulus, or that quiescence, which mark our 
hope that the present enjoyment will continue, and 
those contractions, withdrawals, or merely random 
movements, by which we attempt to avoid or alter 
the unpleasant experience. Such primitive attitudes 
are hardly judgments; they are, rather, unconsid- 
ered responses. Yet already they concern relation of 
the present to a possible future as much as the qual- 
ity of the present itself. Such primitive attitudes are 
replaced by more complex ones, deliberate seeking 
and intelligent avoiding, only when the immediate 
apprehension of value is coupled with an instru- 
mental cognition which penetrates the temporal 

*Sec p. 13, footnote. 


processus of experience and consciously predicts. It 
is the residence of value in certain experiential con- 
texts, and the relation between the supervenience of 
such experiences and the modes of our possible ac- 
tivity, which is the object of the judgment. 

It may be said that there are two questions : What 
is the value of the experience? and How is such ex- 
perience to be got or avoided? But if it be claimed 
that only the former is a question of esthetics, I 
shall beg leave to differ. The former question is 
hardly a real one, because it answers itself. For 
given experience, there can be no doubt about it; 
and for any other experience the question of value 
is the question whether some immediately apprehen- 
sible (value) quale will accrue in a certain empirical 
context. The object of appraisal is some thing or 
situation as a matrix for experience of a certain 
quality. It is only so that the esthetic judgment can 
have significance for practice, and become a guide 
to art or life. 

In the end, of course, all judgments of value must 
depend on some direct intuition of value. And I 
should suppose that judgments of comparative value 
of belter and worse must depend at bottom upon 
some direct comparison of quality as immediate. We 
are here at a point where discussion might easily 
degenerate into a quarrel about words. All I wish 
to point out is that when what is evaluated is pre- 
cisely and solely what is immediately intuited some 
aspect of this experience itself, or some two items 
of experience, both directly given then this evalua- 
tion is not a judgment in the sense of being some- 
thing which needs verification or could be verified ; it 
is no object of learning or of reasonable discussion. 
Direct valuing and direct preferring, of the immedi- 
ately given as such, are not, I should suppose, mat- 
ters concerning which there could be either mistake 
or argument, though certainly they are essential. 

It may be objected that certain experiences pall 


while others are likely to become increasingly satis- 
factory or valuable. But, of course, given unique ex- 
perience never becomes anything but past. It is the 
continuing object, or certain empirical contexts as 
designated by other qualities than value, which may 
become altered in their value-aspect, in ways which 
the esthetician and the artist are concerned to study. 
Esthetics seeks to disclose the enduring values in 
terms of those things or contexts in which positive 
esthetic quality resides. 

Thus so far as esthetic experience is a judgment* 
or gives rise to knowledge, in the sense in which the 
content of knowledge can be learned or verified, that 
judgment is concerned with the same sort of rela- 
tions between different "givens" in the process of ex- 
perience which figure also in other types of empirical 
cognition. There is knowledge here, not so far as the 
esthetic experience is pure esthesis or coalesces with 
the given, but precisely so far as it transcends the 
given and reaches out to further possible experience 
related to the present in certain ways. 

Nevertheless, the terminus ad quern of the esthetic 
judgment is different from that of merely instru- 
mental or practical judgments, of the type which 
figure in natural science, etc., which are what is more 
frequently meant by the word "knowledge." All em- 
pirical judgment concerns the processus of experi- 
ence. But if I may be permitted the expressions 
the practical judgment asks, "Where do we go from 
here?" while the esthetic judgment asks, "What is 
the use of going ?" There is no such thing as "es- 
thetic experience," since all experience has both the 
aspect of value and the aspect of sign: it is at once 
esthesis and cognitively significant. But there are, 
for any presentation, these two questions : "What is 
the object which is here presented, and how is it re- 
lated to others?" and "What is the value of this 
presented object as an enduring matrix of experi- 
ence?" So far as these concern some temporal span 


which runs beyond the present, both are questions 
the answer to which will be knowledge. For neither, 
is this answer to be discovered merely by sinking our- 
selves in the presentation itself and putting thought 
to sleep. And for neither is it possible to make an 
answer without reference to the specific quality of 
the given experience as pure esthesis. But so far as 
the contrast is between such pure esthesis and con- 
struction or relation by thought with what tran- 
scends the given, pure esthesis is not knowledge or 
judgment. The given as such need not be judged. 

If the mystic and the intuitionist are outraged by 
the appropriation of the term "knowledge" to such 
narrow use, one would gladly give them back their 
beloved word, if only they will provide us with some 
other which will mark the distinction which is it nec- 
essary to make here the distinction between the im- 
mediately apprehended and incommunicable "f eel" or 
sensuous quality, which may have intrinsic value and 
be an end in itself, and the instrumental or prag- 
matic significance of other-than-itself which attaches 
to this given, and whose importance lies in its ex- 
trinsic or leading character (its meaning as a sign) 
and in its connection with our possible action. 



It has been pointed out in Chapter III that the 
concept, as that term is here used, represents an ab- 
straction from the cognitive mental state, which in- 
cludes also an element of esthesis representing, for 
the individual mind, the denotation or application 
of this concept in terms of the given. 

It will appear, from Appendix B, that for the sig- 
nificance of a particular mental state as apprehen- 
sion of value, this element of esthesis may be decisive. 
So far as experience can have the aspect of end in 
itself, the determination of value cannot be inde- 
pendent of the immediately apprehended quality of 
the given. But also it will be evident that for evalua- 
tive apprehension of what is objective of objects, 
situations, and typical configurations of possible 
experience the conceptual element is likewise indis- 
pensable. If, for example, I wish so to direct my con- 
duct as to secure a maximum realization of the posi- 
tively esthetic, then I must be regardful not only of 
the immediate qualities of such experiences as will be 
possible to me, but also of those connections of one 
experience with another which constitute the essence 
of conceptual understanding and make it predictive 
and hence a guide to action. 

There are, moreover, interests which transcend 
those of knowledge with regard to other people. So 
far as I am interested in other persons only to se- 
cure their cooperative behavior for purposes of my 
own, my interest is purely cognitive. It ends in their 
behavior and is regardless of the quality with which 
I suppose our common experiences are felt by them. 
But so far as I desire, or feel it my duty, to treat 
other persons not only as behaving objects present 



In my environment but also as ends in themselves, I 
have an interest, which cannot be abstracted from, 
in the absolute quality of their immediate experience. 
The importance of such interests is commensurate 
with the significance of love and duty in human life. 
So far as we have such social ends and purposes, 
in distinction from aims which are exclusively indi- 
vidual, we are under the necessity to frame notions, 
and to seek truths, with respect to which the quality 
of experience in another mind is determinative. The 
ethical conception of the good, and the esthetic con- 
ception of the beautiful, are such notions, if we take 
these to be social categories and suppose that indi- 
vidual appreciations ought to be subordinated to, or 
at least regardful of, prevailing or general appre- 
hensions of the valuable. For instance, it will be of 
the essence of the "good" as a predicate of objects 
in general that it be a possible matrix of satisfac- 
tory experience, not only to me but to the generality 
of persons. In this, it stands in contrast to a purely 
conceptual community of meaning. As was pointed 
out in Chapter II, if I should, by some idiosyncrasy 
of sense, apprehend "red" by immediately given 
qualia peculiar to me, but should, on account of the 
social origins of language, apply this term to the 
same objects as other persons that is, find it in the 
same patterns of relation then my cognitive con- 
cept of red would be identical with that of other 
persons, regardless of its peculiar quality as im- 
mediately given. But if my meaning of the "good" 
should represent a similar idiosyncrasy, then the 
purposes for the sake of which I have framed and 
use this term would be defeated, whether I should 
know it or not. I do not mean to designate as 
"good" what other persons merely behave toward in 
the same way I do or find in the same contexts as I 
do : I intend by it that which affects them with the 
same, or similar, qualities of experience with which 
I am affected in the presence of it. Notions so 


framed may appropriately be termed "ideas. ** Such 
ideas are basic for the sciences of values ; for ethics, 
esthetics, and the philosophy of religion. In terms of 
them all truths of (social or objective) appreciation 
must be framed. 

The directly given quality of experience, however, 
can never be conveyed or expressed; it is ineffable 
and incommunicable. I seek to interpret other minds 
by empathy or einfuhlung, and inference from be- 
havior. That such sympathetic comprehension is not 
pathetic fallacy, there is so far as I can see no 
theoretical assurance. It transcends the possibility 
of verifiable knowledge, and can be founded only on 
a postulate. 

If the taxi-man has a proper concept of an hour 
and can meet me at the end of that time, we achieve 
a perfect practical understanding regardless of the 
tedium or vivacity with which the passing time may 
differently affect us. The hour we are concerned 
about, in our interests of cooperative behavior, is a 
purely relational thing, which may be the same for 
both of us regardless of the fact that, if I could put 
myself inside his experience, this hour might feel 
twice as long as to me now. For me to disregard this 
immediate quale which the time has for the taxi- 
man, is to treat him as a thing and not as a fellow 
creature. But I cannot literally Jcnow his experi- 
ence : I can only postulate that his behavior is a clue 
to it. Upon such postulate, ethics must be founded. 

There are many motives among them a powerful 
and perhaps central motive of religion which move 
us to extend such postulation to reality as a whole. 
We are interested not only to manipulate things and 
predict their behavior, but to question whether re- 
ality is stable in a fashion which is not merely the 
reliability of obedience to natural law. Has reality 
on the whole a stable relation to our values ? Does it 
conserve the humanly ideal, or does it inevitably 
frustrate us in the end, or is it "indifferent" to our 


valuings? We inevitably raise this question. To sup- 
pose that it means anything or has an answer, is 
inevitably to attribute to what we regard as ulti- 
mate reality one essential character of persons de- 
termination by relation to value. But if "ultimate 
reality" means anything; if, further, ultimate real- 
ity has an "attitude" toward human ends if it be 
"personal" ; still I can only read this mind of God 
by a postulate, however much the vehemence of faith 
and the yearning to break through that loneliness 
which is the fate of self-conscious beings may move 
me to affirm it as knowledge. 

Apparently it is a native longing of humanity to 
transcend the bounds of subjectivity; to know our 
object not only in the pragmatic sense of successful 
prediction and control but in a deeper sense of 
somehow coinciding with its nature. To represent 
this coincidence of subject and object as genuinely 
possible and as true knowledge, is characteristic of 
mysticism, of intuitionism, and of that type of ideal- 
ism represented by Schopenhauer and Bergson. The 
world is my idea says Schopenhauer but the no- 
tion that it is nothing more than my idea is incredi- 
ble. That would be the solipsism which is maintained 
only in mad-houses. But what more than my idea can 
it be, consistently with my genuine knowledge of it? 
If it is something more than what it means for me, 
something in itself, then it must mean something for 
itself ; it must, in this respect, be of a nature funda- 
mentally like my own. Insight into the true nature 
of a reality which is independent of me which has 
more than a "for me" character is possible only if 
that nature is spiritual. My immediate experience is 
clue to it only because, in its character of will, my 
nature coincides with the nature of all reality. The 
parallel, for other and similar conceptions, the 
reader will be able to draw for himself. 

In brief, such idealism, as distinguished from other 
types by its admission that reality is independent of 


the mind which knows it, rests upon a dilemma which 
is real and is, by such idealism, correctly understood: 
Either knowledge does not mean identity of quality 
or nature between subject and object, or the only 
intelligible fashion in which reality in general can be 
conceived is on some analogy to mind or life, as 
spiritual. Such idealism chooses the latter alterna- 
tive. As I have tried to argue, the other is the true 
one. In the case of other conscious beings, empathy 
has a meaning. In the case of the inanimate, it is 
dubious whether such meaning is possible at all. But 
in any case, knowledge is independent of any sup- 
posed identity of quality between subjective know- 
ing state and objective thing. Other interests may 
concern such identity, but knowledge does not. Gen- 
uinely verifiable knowledge cannot, thus, interpret 
things on any analogy to spirits* On the contrary, 
it can grasp other minds only as things, revealed in 
the patterns of behavior of certain physical beings. 
The rest is postulate. 



It will very likely appear to the reader of Chap- 
ter II that the analysis of the cognitive experience 
there given leads to a difficulty about our knowledge 
of the mind itself. This difficulty is, I think, more 
apparent than real, and is mainly due to that back- 
ground of our thought for which the history of 
modern epistemological developments is responsible. 
Perhaps the matter may be briefly put as follows : 
Compatible with conceptions here presented, the 
ascription of reality to anything as an object of 
knowledge must represent an interpretation put upon 
some presentation in experience. But the mind, too, 
must be knowable. Either that, or it must be ac- 
cepted as transcendent of experience, which would 
make of it just that kind of dubious and metaphysi- 
cal assumption which has here been declared against 
in the theory of knowledge. Apparently, then, the 
mind itself must somehow be given. Mind, however, 
is exactly that to which the element in experience 
which is TIG* given has been ascribed. If, then, the 
mind and its activity are given, then everything in 
cognition is given, and the distinction fails. But if 
mind is not given, then how can it be known; how 
can it be other than a meaningless fiction of the 

The nature of mind in general is a fundamental 
and complex problem of metaphysics, including 
much with which we are not here concerned. While 
it is not possible to avoid such metaphysical prob- 
lems altogether, the attempt will be to restrict dis- 
cussion directly to the point in hand that is, to 


explain the possibility of the mind's knowledge of 
itself, compatible with the conception that the knowl- 
edge of a real object is through interpretation of 
something given in experience, and that the element 
of interpretation in knowledge is due to mind. 

The key to this problem may be discovered, I 
think, by a slightly critical examination of what is 
involved in our ordinary ascriptions of reality and 
our usual attributions of "responsibility." As was 
pointed out in Chapter I, a proper method in phi- 
losophy will take it to consist in just such critical 
clarification of common modes of predication. What 
follows, then, attempts merely to draw attention to 
those phases of experience which are what we mean 
when we say that thinking qualifies the object of 
knowledge, and when we say that the mind itself is 

My experience of the pen with which I am writing 
has been analyzed into a presentation now before 
me and an intellectual construction or interpreta- 
tion put upon it. When I say that this given cylin- 
drical appearance is due to the real fountain pen in 
my hand, I mean that, as I feel assured from past 
experience, there are various continuities of this 
given presentation with other actual and possible 
experience in ways which are characteristic of just 
this kind of real physical object. The givenness of 
just this kind of appearance, together with just 
these specific continuities, constitute the meaning of 
"a physically real fountain pen now in my hand." 
Similarly when I say that the taking of it as "pen 5 ' 
and not as "plaything 9 * or merely "cylinder" is due 
to my mind, I mean that, as I feel assured from past 
experience, if my attitude and purpose were other 
than they are, the significant cognitive context of this 
presentation would be other than it is. The elucida- 
tion of just this correlation between different mental 
attitudes and interests, on the one hand, and different 
significant contexts of the given on the other, would 


be the explication (or partial explication) of the 
meaning of "my thinking mind conditioning the con- 
tent of experience. 9 * 

It might be said that this is roughly comparable 
to a more naive pronouncement, which might be at- 
tributed to common sense : There is a brute-fact ele- 
ment in knowledge which is distinguished as that 
which is due to the presence of the object; and there 
is an element of relation, association, or construc- 
tion which is distinguished as being due to the mind. 
If, however, we should suppose that actually being 
"due to the object" and being "due to the mind" 
are the criteria which mark this distinction between 
elements in cognition, then we should have the cart 
before the horse. We should then be committing the 
error, which is frequent in naive realism, of sup- 
posing that a distinction within knowledge can de- 
pend on a prior one outside it. To be sure, any plau- 
sible analysis of knowledge must be consistent with 
the statement that the veridical perception of the 
object in the mind is due (in part) to the presence 
of the object to the mind. Even idealism must find 
some eventually valid meaning for ordinary pro- 
nouncements about the causation of perceptive 
states. Still, being due to the object is no criterion 
of givenness, because the recognition of the presence 
is itself dependent on this givenness of the presenta- 
tion. The presentation is due to, or caused by, the 
object; but knowledge of the object is due to the 

An exactly similar difficulty would hold for any 
theory which should attempt to delimit or explain 
the "formal" element of interpretation by the cri- 
terion of its being due to the mind. That such con- 
struction is thus due to mind may be accepted as 
a fact; obviously something in knowledge will be 
due to mind if there is any such thing as a mind. 
But just as the object can be known only through 
or by means of that presentation which is due to it, 


so mind (as cognitive) can only be known through 
that "formal" element in experience to which it 
gives rise. If we did not find both the given and in- 
terpretation in experience, we should have no epis- 
temological ground for the distinction of subject 
and object. Hence to rest the division of cognitive 
experience into "matter" and "form" upon a sup- 
posedly prior distinction of mind and independent 
object, is to commit an obvious and vicious circle. 

It is necessary however undesirable to pause 
upon this methodological point, because it is pre- 
cisely what lies at the root of the supposed difficulty 
about conceiving that a mind which is one of the 
conditions of experience in general should be itself 
knowable and empirical. The alternative is tran- 
scendentalism, which explains knowledge by refer- 
ence to a conditioning mind and to avoid the para- 
dox of mind conditioning itself posits this mind as 
beyond or behind all experience. Such transcenden- 
talism vitiates both epistemology and metaphysics. 
It vitiates epistemology because to suppose that the 
content of knowledge is informed or determined by 
some transcendent agent or principle, which is not 
to be found in experience, is to substitute a creation- 
myth for the analysis of knowledge. And it vitiates 
metaphysics by this invention of a transcendent 
agent which, on the very account of knowledge 
which such a theory gives, it would be impossible 
to know. The ascription of the categories to a tran- 
scendent mind and of the matter of knowledge to 
transcendent things in themselves, are precisely simi- 
lar fallacies. Those who hope to avoid the latter 
must, in all consistency, give up the former. 

It may be said that mind, as that which supplies 
the categorial conditions of experience, is something 
which, in the nature of the case, must be beyond or 
behind experience and cannot be in it, but that the 
existence of such a mind is "presupposed" by ex- 
perience in general. If this vague word "presuppose" 


has any real and pertinent meaning, I should sup- 
pose its connotation would be of an hypothesis or 
assumption which is logically necessary to explain 
the facts in question. But the difference between an 
hypothesis which is absolutely required in order to 
explain a fact, and direct experience of the thing 
hypostatized, is a wholly imaginary difference. To 
see this, we should begin, not with hypotheses which 
are required or necessary but with such as are only 
more or less probable. That which explains experi- 
ence is always something which the experience in 
question gives us some reason (some partial ground) 
for assuming. So put, this may seem a commonplace. 
But if this point be fully grasped, it will throw much 
light upon the connection between the probability 
of hypotheses and the "perception" of "empirical 
facts." It will also do much to resolve such difficul- 
ties in the theory of knowledge as the one with which 
we are here concerned. 

Let us return to our illustration. At the moment, 
the content of this present experience leads me to 
say that I have a fountain pen in my hand. That is, 
I explain or interpret certain data of sense by the 
hypothesis of the fountain pen. This hypothesis is 
partially verified by the presented data of sense 
themselves. For the rest, it depends upon the verifi- 
able truth of certain implicit predictions which con- 
stitute my interpretation or explanation of these 
sense-data. For instance, I expect it to continue to 
write at my will, or if it does not, then I expect it 
to begin again after I dip it in the ink-bottle and 
manipulate the lever. These implicit predictions 
too numerous and complex to mention are all of 
them about further possible experience. If experience 
could exhaust all such prediction, what I mean by 
"the fountain pen in my hand" would completely 
coincide with what this totality of experience would 
include. When I say\that this present content of my 
experience is "due to" the fountain pen, I state a 


highly probable fact. When I say that the fountain 
pen as a real and knowable object is an interpreta- 
tion that I put upon my present experience by 
framing this supposedly verifiable hypothesis about 
further possible experience I state the same fact 
in another way. 

Between the fountain pen in my hand, the dis- 
tant star, and an alpha-particle or electron, there 
is no difference in type, from the point of view of 
my knowledge, but only a difference of degree in 
the probability of my interpretation, or the possible 
discrepancy between it and what further experience 
might disclose. And whatever, in my interpretation, 
is thus probable only, is intrinsically capable of 
being verified or falsified in such possible experi- 
ence. There is nothing in it which is beyond the pos- 
sibility of experience altogether. If there were, it 
would also be beyond all meaning. The difference of 
these examples is in the extent to which what is 
meant by "the pen," "the star," "the electron," is 
verified in the presently given sense-data, or con- 
versely, in the degree to which what is meant is not 
thus immediately verified and hence is only more or 
less probable. The pen is "directly," though only 
partially, given at the moment. The star of alpha- 
particle is "indirectly" given; that is to say, be- 
tween the immediately presented point of light or 
the track in the photograph and the full meaning 
of "a star" or "an alpha-particle," there is a more 
obvious or greater interval. The electron, I may be 
tempted to say, nobody can perceive. But in experi- 
encing those laboratory phenomena which oblige us 
if they do to believe in the existence of elec- 
trons, the electron is partially given. This behavior 
of the oil droplet, or whatever, requires the exist- 
ence of electrons to explain or interpret it. And the 
only kind of reality I can ascribe to the electron, is 
the nature I attribute to it in explaining this and 
other pertinent experience. As Charles Peirce has 


phrased it: "Consider what effects that might con- 
ceivably have practical bearings you conceive the 
objects of your conception to have. Then, your con- 
ception of those effects is the whole of your concep- 
tion of the object."* 

Apply these considerations to the conception of 
a supposedly transcendent mind or agent which cate- 
gorizes experience. If there were nothing in experi- 
ence which was the datum of mind, there would be 
no possible ground for assuming the existence of 
any. If, on the other hand, the only rational ex- 
planation of certain features of experience is the 
hypothesis of mind having a certain character or 
exercising certain functions, then this logically ne- 
cessitated explanation of such experience would 
constitute our knowledge of such mind. A mind 
which is "presupposed" or necessarily assumed in 
order to account for experience is partially revealed 
in every experience which thus "presupposes" it, 
and the whole meaning of it would be exhausted in 
the totality of such experience. There is nothing in 
the nature of it which can be meaningfully asserted 
beyond what such experience, actual and possible, 
would exhibit. So far as it is beyond all possible 
experience, it is not a conceivable explanation of 
anything. So far as it is even a possible explanation 
of experience, it is not transcendent of experience 
in general. The possible transcendence of mind, ex- 
actly like the transcendence o the real object, is the 
interval between what in our conception of this 
mind present experience verifies and what nothing 
short of the totality of pertinent experience would 

We are now in position to approach the heart of 
the difficulty about the "givenness" of mind. Two 
elements have been distinguished in the cognitive ex- 
perience; the given, which is characterized by our 
inability to remove or alter it merely by an activity 

""Chance, Love, and Logic.** p. 45- 


of thought, and a construction or interpretation 
which is attributed to the activity of mind. But this 
mind must be identifiable through some sort of datum 
of experience and must have its meaning in the em- 
pirical in general. 

The activity of mind is evidenced, first, in the 
feeling of such activity. We should most of us have 
difficulty in describing this feeling, because we are 
normally inattentive to those correlates of it which 
can be stated otherwise than in terms of its effects 
outside the body. Thus we may be quite unable to 
say whether what we call the feeling of thinking is 
a feeling of innervation or certain specific cona- 
tions or only a warmth of the jaw. But though we are 
not here concerned with the psychological descrip- 
tion of the thought process, it may be well to point 
out that the possibility of such description, or even 
of the problem, depends upon the identification -first 
of that whose correlations, or "description" or 
"analysis," is to be given. If, then, the traditional 
difficulty of this psychological problem should tempt 
any one to deny the existence of any mental phe- 
nomenon truly describable as thinking, it may be 
that he commits a fallacy of oversophistication, and 
is looking under the table for what is on it. Let him 
describe what happens when we think we think, and 
if his description is a true one, then what he de- 
scribes is what ought to be called thinking. But 
whoever should be unacquainted with the feeling of 
thinking would automatically be precluded from at- 
taching any meaning to the term at all. 

The effect of thinking, we learn as we learn the 
effect of flexing our muscles or of any other activ- 
ity. When I interpret this thing in my hand as "a 
cylinder,' 5 its shape forthwith stands out in the field 
of attention and other aspects of it relatively fall 
away; the relation of it to one set of other things 
(seen or imaged) rise before me, and other such re- 
lations may lapse. The image of it rolling off my 


desk may come into consciousness, whereas the mem- 
ory of paying money on receiving it probably will 
not. I learn that what is thus classed under the head 
of construction or interpretation is due to my ac- 
tivity much as I learn that the shape of these 
scratches on the paper is due to my activity. If I 
purpose to write, the scratches come; if I purpose 
to wait until I am clearer what I want to say, then 
for a time no more scratches. In general, I learn 
my own activity through the correlation between 
certain directly observable feelings, classed by com- 
mon sense as desire, interest, purpose, etc. 5 and cer- 
tain externally observable happenings, (These feel- 
ings are as much given items of experience as the 
feeling of presented red or soft. The correlation of 
them with such further items as the externally ob- 
servable happenings is not, In the same sense, given 
but is a generalization the truth of which we learn 
by an induction from many such experiences. The 
feelings are conceived or named 9 as things in gen- 
eral are conceived and named, in the light of such 
learned stable relationships* Thus such a category 
as "purpose" or even "attention" may be psycho- 
logically inept, because it may be the case that a 
correlation between these feelings and sufficiently 
specific or fundamental neuroses is difficulty or im- 
possible to discover. But if this should be so, still 
it would be no criticism of common sense nor of 
the use of such categories in epistemology, because 
both common sense and the analysis of cognition 
are relatively little concerned with brain and ner- 
vous processes, or even with those relatively slight 
and subtle bodily happenings frequently connoted 
in psychology by "behavior," For example, I am 
familiar with the feeling of intending to write. The 
correlation between this and the scratches on the 
paper is likewise familiar. But I do not know what 
muscles I use when I write, to say nothing of in- 
tervening processes less easily observed. Nor does 


an interest in epistemology dictate that I should be- 
come specially interested in them.) 

In the course of experience, I learn the correla- 
tion which exists between such feelings as attending, 
concentrating on a given item, being interested, and 
the further associations or configurations which 
then accrue or stand out in consciousness. If it be 
asked, "How do I know that it is the mind to which 
such characteristic alterations are due? 55 the an- 
swer is that I learn what is due to the mind by the 
difference which it makes if I refuse to attend or 
am differently interested in the given. For the rest, 
the answer is indicated by the above observations 
concerning the relation between the phenomena ex- 
plained and that which they are explained as "due 
to.' 5 What I mean by the mind is partially revealed 
in just such feelings of purpose, desire, interest, and 
the like, and these other alterations in the proc- 
ess of experience which are attendant upon them. 
Whether this mind is an immaterial agent, or a 
Democritean complexus of smooth round atoms, is 
simply a further question with which we are not 
here concerned. "The mind 55 in general includes very 
many other kinds of phenomena beside those of cog- 
nition. To characterize the mind as a whole is a 
metaphysical problem of the first magnitude. But 
on any metaphysical account which should be even 
conceivably correct, what we mean by mind must 
be just what is revealed in the totality of those 
phenomena which are ascribed to mind. To try to 
find the mind in any other sense, is the counterpart 
of the fallacy of attempting to discover the sub- 
stance of the physical object apart from all its 
qualities, relations, and effects. What is meant by 
"the activity of thinking" is precisely such correla- 
tion between attending, reflecting, etc., and those 
alterations of context of the thing attended which 
experience shows will ordinarily and normally be 
induced by these. We learn the existence and nature 


of thought, as we learn the nature and existence of 
anything else, through the difference that it makes 
in experience. The mind and its activity transcend, 
of course, what is revealed in any particular case. 
Ascribing the particular phenomenon to mind is an 
interpretation; the significance of what is predi- 
cated is not exhausted by this phenomenon itself 
but includes a relation of this to a multitude of 
others, which the predication implicitly asserts to 
be intrinsically possible of experience. Such a predi- 
cation, being interpretive, can be mistaken in the 
particular instance. But the totality of such valid 
interpretations or attributions, exhibit the mean- 
ing of "the cognitive activity of mind." 

I believe that we may now remove the last major 
difficulty about our knowledge of the mind if we 
observe that although the activity of mind is a 
datum of experience, it is not the kind of datum to 
which the word "given" has been applied in the jpre- 
ceding. A comparison may be helpful here. It is a 
datum of experience that I can move my arm, or 
that I can twitch a certain muscle at will. But this 
fact is something which I have learned; it repre- 
sents a generalization from experience, not the mere 
givenness of the twitch. This fact or generalization 
is the correlation between the twitch and my inten- 
tion. That this correlation exists, is a datum of ex- 
perience an unalterable fact but this general or 
stable fact is not given with the twitch. It is because 
of this reliable correlation, which I have learned, 
that I call the muscle-twitch something that I do. 

Similarly, the ascription of certain phases of the 
cognitive experience, here called "interpretation," to 
the mind, represents the empirical generalization 
that between this element in experience and our at- 
titudes of attending and reflecting there is a reliable 
correlation. The existence of this correlation is an 
absolute datum; but it is not the kind of datum 
which "given" is defined to mean. This datum is the 


fact that when we thus think, then certain contexts 
of the thing thought about accrue, in ways in which 
they do not accrue when we think to some different 
purpose or fail to attend or reflect. When I think 
this in my hand as "cylinder," the image of rolling 
adds itself: when I think it as "a poor buy," or 
when I am thinking what I want to write and so 
ignore this item in my field of vision, the image of 
rolling is absent. This Image, taken by itself, would 
be homogeneous with "the given," just a voluntary 
muscle-twitch, by itself, is homogeneous with an in- 
voluntary twitch. But the one twitch as I have 
learned is correlated with my intention, while the 
other is independent of it. Similarly, the visual pat- 
tern of the pen in my hand and the image of it roll- 
ing are both data of experience, under certain cir- 
cumstances. But it is. the circumstances which mat- 
ter. Whether I choose to find this visual pattern as 
it now is or some other shape, it remains just this. 
That item answers to the criterion of givenness: it 
is independent of my attitude of thought. But the 
accrual or non-accrual of the image of rolling and 
other such context is as I have also learned not 
thus independent, but is correlated with my attitude 
of thought. This context, then, answers to the cri- 
terion of construction or interpretation, of being 
"due to the mind," To say that the activity of mind 
is "given," would thus be subtly incorrect. But to 
say that the activity of mind is something known 
through certain data of experience data which 
are in one sense homogeneous with those designated 
as "given" is no paradox but, I should suppose, a 
more or less obvious fact. The mind, as known, rep- 
resents our interpretation of such data; their pro- 
jected relation to a multitude of other such instances 
and to further experience which we take to be intrin- 
sically possible. 

The paradox of mind knowing itself is merely one 
of language. If it appears to be comparable to the 


puppy trying to turn around fast enough to catch 
his own tail, that is because it is falsely supposed 
that the thing to be known is identical with this act 
of knowing. The mind which is known transcends 
the momentary knowing just as the external object 
known transcends its instantaneous phenomenal ap- 
pearance. I know my mind by means of generaliza- 
tion from certain phases and aspects of past experi- 
ence which are, by construction, added as context to 
similar phases of my present experience; just as I 
know the object by means of generalization from 
past experience which, by interpretation, is added 
as context to the present phenomenal appearance 
of it. And that present datum of experience which is 
interpreted as "activity of thought" is just as ob- 
jective and intrinsically observable a kind of datum 
as is the phenomenal appearance of an external ob- 

The mind and particularly its purpose and ac- 
tivity is, of course, ultimately mysterious, just as 
concentration upon the presentation of the starry 
heaven reveals it as something ultimately mysterious, 
when all those prosaic and familiar correlations of 
this and that, which constitute its explanation, are 
shorn away, and we stand before it in its pristine 
glory. But the mind is mysterious in no specially 
different sense than this one in which reality alto- 
gether is ultimately mysterious. 

The conception of any esoteric or peculiar type 
of knowledge by which we grasp the nature of our 
own minds, is both implausible and unnecessary. 
That the mind is at once that to which interpreta- 
tion of the given in experience is due, and something 
itself known by interpretation of certain characteris- 
tics or data of experience, is as has often been ob- 
served no more paradoxical than that we should 
be able to change our own position as well as the 
position of other things, or direct any other activ- 
ity upon ourselves. 


The feeling of paradox which may persist here is 
a legacy of the transcendentalist mode of thought. 
If mind he conceived as an ultimate reality, while 
that which results from the interpretation or con- 
struction by mind is conceived to have a merely 
apparent or lower order of existence, then obviously 
the real mind cannot know itself. It is for just this 
reason that the transcendentalisms elaborate story 
of the categorizing of experience by the mind is in- 
consistent with itself, since if this account of knowl- 
edge should be true it could not be known to be true. 
Similarly if what is "known as" an interpretation 
by the mind is created by or "constituted" by that 
interpretation, then the mind cannot know itself, 
since we can hardly imagine what we should mean 
by saying that it created itself or constituted itself. 
Fichte, of course, is driven to just this desperate 
strait; the mind "posits itself, and this is the first 
act of positing." Thus mind becomes the ultimate 
and hopelessly esoteric mystery* 

In any discussion of mind in the theory of knowl- 
edge, it is particularly desirable to observe that 
there are two characteristic modes of "explanation" 
by the ratio essendi and by the ratio cognoscendi 
and that these two run in opposite directions. In 
the case of the external object, this is clear enough 
and prosaic enough. The ratio essendi of the sense- 
data is the presence of the real object; and the ratio 
cognoscendi of the object is the sense-data them- 
selves. Likewise in the case of the mind; the real 
mind is the ratio essendi (the cause of) the inter- 
pretational or "formal" aspects of cognitive expe- 
rience, and these aspects are themselves the ratio 
cognoscendi (the clue to) the cognizing mind. The 
case is not fundamentally different than for real 
things and their appearances or phenomenal effects 
in general. The star is the cause of the light and the 
light is evidence of the star ; the electron is cause of 


the motion of the oil-droplet and this motion is evi- 
dence of the electron. At the limit, the nature of the 
real cause coincides with the totality of its conceiv- 
ably experienceable effects. But this does not remove 
the disparity between the real nature of the cause 
and the presently observed and meager revelation 
of it. Nor does it invalidate the different significance 
of the two opposed directions which our characteris- 
tic modes of understanding may take. The causal or 
"cosmic" explanation, of science and common sense, 
runs from cause to effect, from hypostatized thing 
to observable evidence. The analysis and verification 
of knowledge runs from effect to cause, from evi- 
dence to the thing evidenced. Perhaps it needs to be 
emphasized that one type of explanation or analysis, 
satisfying certain purposes, does not exclude or 
make superfluous other types of explanation,, insti- 
gated by a different interest or made from some 
other point of view that one analysis may be 
wholly true without being the whole of the truth. 

Epistemological investigation is, naturally, by 
way of the ratio cognoscendi; that is its peculiar 
task. Those "theories of knowledge" which reverse 
the direction of explanation and give a causal, nat- 
ural-scientific account, merely substitute a more or 
less uncritical and psychological methodology, based 
upon dubious assumptions, for their proper business. 
Transcendentalism is, in general, the result of the 
opposite fallacy of attempting to base everything 
on the theory of knowledge. It tries to suppress, as 
superfluous or merely secondary, all natural-scien- 
tific explanation of those phenomena of which it 
takes cognizance, and to substitute for it an analy- 
sis of our knowledge of these phemonena. Thus it 
ends by identifying cognition and creation, by af- 
firming that there is no ratio essendi save the ratio 
cognoscendi, the content of knowledge or the mind. 
The characteristic result is the reduction of the 


reality which appears to the appearance itself, and 
the elevation of the mind above appearance to a 
realm in which it is not knowable as other things are 



There is one problem about the application of ab- 
stract conceptual systems to experience, not men- 
tioned in Chapter IX, which merits a little atten- 
tion. That is the question: How far does the char- 
acter of given experience determine the applicabil- 
ity or inapplicability of concepts, when these con- 
cepts are first envisaged as abstract patterns of 
purely logical relations? In the first place, it is ob- 
vious that if we take a conceptual system com- 
pletely in the abstract, as is done in the case of pure 
mathematics, there is no indication in the system 
itself of its applicability or inapplicability to any- 
thing. It may be remarked, however, that the very 
fact of its being developed at all is evidence of its 
applicability to something. I do not mean by this 
that such systems of mathematics originate as ap- 
plied or empirical truth, and are later abstracted 
from such application. That is true, in general, 
though there are such exceptions as quaternions and 
the non-Euclidean geometries. What I mean is, that 
whether "imageless thought" is psychologically pos- 
sible at all or not, no human being has, or ever will 
have, logical powers sufficient to enable him to elabo- 
rate the analysis of concepts in systematic fashion 
without reliance upon imagery. If the mathemati- 
cal system applies to nothing else, it will apply to a 
set of distinguishable arrangements of symbols used 
according to certain rules. This is a relatively trivial 
application to experience, but it is an application to 
something empirical nevertheless. In fact, we might 
well ask if the designation "abstract*' as applied to 



mathematics is not a figure of speech, since no sys- 
tem is to be discovered, in the mathematician's mind 
or elsewhere, in complete isolation from denotations. 
The point is, however, that such denotations may be 
"accidental" ; and there is no such denotation from 
which the system cannot be separated (and trans- 
ferred to some other) while retaining its intended 
identity as a system of abstract concepts. 

The well-known practices of mathematics, in de- 
vising tests of consistency and independence, will at 
once confirm in the reader's mind the fact that any 
abstract conceptual system, however elaborate, will 
be such that more than one empirical denotation 
could be found for it. It represents a type of order 
which is exhibited by more than one kind of empiri- 
cal things. Since this is true of mathematical sys- 
tems, which are relatively complex, it is fairly 
evident without further corroboration that it will 
also be true of concepts in general. 

It is not quite so clear whether the same empirical 
content may be denoted by different conceptual sys- 
tems different not in name simply but in that pat- 
tern of logical order which is the essence of the ab- 
stract concept. Out of hand, one might suppose that 
this will not be possible unless comparing two sys- 
tems of concepts to be applied to the same content 
the application of one set of concepts neglects 
certain items of the empirical which the other set in- 
cludes in its denotation. In general, that is probably 
true. If two conceptual interpretations should both 
be applicable to the same content, the one will be 
more abstract or general than the other, or will be 
abstract in a different way. In that sense, the appli- 
cability of different concepts to the same empirical 
thing or situation is a commonplace, since there is 
nothing which can be named by one name which can 
not also be named by some other desk, furniture, 
convenience, wood, antique, expensive, ugly! It is 
more to the point to observe that different abstract 


conceptual systems which are quite complex and are 
of the same order of structural complication may 
be applied to the same empirical content and include 
in their denotation the same empirical characters, 
though they are quite different conceptual systems 
as types of logical order. A good illustration is that 
kind of point-for-point correlation in denotation by 
which it may be proved that if Euclidean geometry 
is free from inconsistency, then some non-Euclidean 
system say Riemann's is so also. One may apply 
Riemannian plane geometry to the surface of a 
hemisphere by letting Riemannian "straight" denote 
Euclidean "great-circle cut orthogonally by a 
plane," etc. An even more illuminating illustration 
would be the systematic reinterpretation of New- 
tonian-Euclidean facts in the kinematics of relativ- 
ity theory. 

Envisaging the abstractly conceptual as a type 
of order or pure pattern of logical relationships, one 
might say that at least the given character of the 
empirical imposes certain limits upon the concepts 
which can be applied to it; that a conceptual sys- 
tem cannot be imposed if it requires distinction 
where no distinction in the given can be marked, or 
relation where no conjunction in the given can be 
found. Such a statement would at least draw atten- 
tion to an important consideration which seems to 
limit the usefulness of applying a conceptual inter- 
pretation to a given content. In general, a con- 
ceptual interpretation which makes distinctions or 
relations where, in experience, none are to be discov- 
ered, will be a poor intellectual instrument for deal- 
ing with that phenomenal content. And yet, there is 
abundant evidence that we not only can but do apply 
concepts to the given where this involves making 
distinctions and relations which empirically cannot 
be discovered. Not only that, but often such inter- 
pretation serves a useful purpose and is highly sci- 
entific. For instance, the array of color-qualities is 


interpreted as a continuum, by correlation with vi- 
bration-frequencies, etc., although the directly dis- 
criminable color-qualities are certainly neither in- 
finite in number nor such that between any two 
another can be found- The same is true of pitches ; 
and exactly this procedure is frequent in the sci- 
entific interpretation of sense-qualities. In fact, one 
might say that arrangement in a one-, two- 5 or n- 
dimensional continuum is the scientific way of at- 
tacking all sorts of areas of experience, though 
no type of the empirical, as given, is genuinely con- 
tinuous in the sense required by such mathematico- 
scientific analysis. Absolutely never is it capable of 
direct differentiation into the infinity of distinct gra- 
dations which characterize the order of the mathe- 
matical continuum. It is for this reason (amongst 
others) that Bergson repudiates such scientific in- 
terpretation as misrepresentative of reality. And 
Professor Whitehead, by his <c method of exten- 
sive abstraction," has shown us how situations of 
this sort can (ideally) be reinterpreted so as to dis- 
close the order which is conceptually required as 
genuinely present, not as an order of the directly 
given but as an order of our own ways of ordering 
it. In fact, although consideration of space and of 
the reader's interest forbid extended illustration, a 
little attention to scientific procedure makes it clear 
that one could pile Ossa upon Pelion of evidence that 
conceptual systems may usefully and validly be ap- 
plied to given content although they involve making 
distinctions which are not given and establishing re- 
lations where none are found. 

To the question, "What abstract concepts or sys- 
tems could be applied to what given content?" the 
only answer which seems possible is: It is unsafe to 
say that any concept could not be applied to any 
empirical content. In this, as in other respects, it is 
hazardous to set bounds to human ingenuity* or 
even to delimit a priori what could be made useful. 


To the question, "What will determine such applica- 
tion?" the answer seems to be: Complex considera- 
tions, in which the purposes to be served, the type of 
order of the concepts, and the general character of 
the given, will all be important, but in which, ap- 
parently, no one of these factors by itself is capable 
of being decisive. This, as it seems to me, but serves 
to emphasize the fact that the conceptual interpreta- 
tion of experience is, at bottom, something concern- 
ing which rationalistic accounts and empiricistic 
theories are, in their opposite ways, both false, and 
the pragmatic is the true one. 



The theory of the a priori set forth in Chapter 
VIII contains two main theses: (1) All a priori 
truths are definitive; they explicate criteria of clas- 
sification, including the criteria of reality in its va- 
rious categories; and conversely, all criteria of 
reality are a priori and independent of experience 
because they concern the classification or interpreta- 
tion of empirical given content and do not indicate 
or determine that content itself. (2) The choice of 
concepts and systems of such for application to ex- 
perience is pragmatically determined. 

Another way of expressing the first of these two 
would be to say that a priori propositions coincide 
with the class of truths which are analytically de- 
termined and with propositions true in intension; 
what is a posteriori coincides with the logically syn- 
thetic and with propositions true in extension. The 
equivalence of the a priori, the analytic, and the in* 
tensional, on the one hand, of the a posteriori, the 
synthetic, and the extensional, on the other, has fre- 
quently been denied. Failure to observe these equiva- 
lences has led to extreme confusion in logic, much 
of which persists at the present time. The most im- 
portant topic in this connection would be the mean- 
ing of implication and the nature of inference. But 
examination of that question would of necessity be 
too long and complex for inclusion here. What can 
be offered in brief space is a sketch of the distinc- 
tion between propositions in intension and in exten- 



The a priori proposition and the empirical gen- 
eralization are usually indistinguishable by their 
form. Both are universal in intent, and are normally 
expressed by an "all" proposition or by one in which 
the "all" though unexpressed is obviously under- 
stood. The difference between these two is that be- 
tween the intensional and the extensional "all." If I 
say, "Parrots are birds," I am correctly understood 
as meaning that any creature not a bird could not 
be a parrot. But if I assert, "Parrots have a raucous 
cry," I should not be correctly interpreted if I were 
taken to mean that a bird with a melodious note 
could not belong to the genus "parrot." The first 
proposition is a priori; the second, an empirical 

The difference between the two may be expressed 
in various equivalent ways. The first explicates the 
intension or essence of the subject-term "parrot"; 
the second asserts only the inclusion of the class of 
actually existing creatures which come under the 
definition of "parrot" in another class, "creatures 
with a raucous cry," The first expresses in the predi- 
cate something logically contained in the subject; 
the subject-concept implies the predicate-concept. 
The second states a factual connection of two classes 
of objects. The first can be transformed into a strict 
hypothetical proposition, "For any X you please, if 
X is a parrot, then necessarily X is a bird.'* But 
if we similarly transform the second, "If X is a par- 
rot, then X has a raucous cry," we must recognize 
that we have here a different meaning of "If . . ., 
then . . ." ; that it does not state a logically neces- 
sary but only a material or factual relation what 
is called by Mr. Russell a "formal implication," 
which holds only of the materially existent, not of 
the all-possible. (Incidentally, this name is highly 
inappropriate.) Finally, we can exhibit this differ- 
ence in unambiguous and decisive fashion if we ex- 
press the meaning of each of these propositions as 


a negation* The first means, "A parrot which is not 
a bird is logically inconceivable" ; the second means 
only, "A parrot without a raucous cry does not ex- 
ist. 5 * Propositions in intension concern what is pos- 
sible, impossible, or necessary; an affirmative uni- 
versal in intension states a necessity and negates a 
possibility. Propositions in extension concern only 
what does or does not exist. It is obvious that a uni- 
versal in intension implies the corresponding uni- 
versal in extension: what is impossible cannot exist. 
But the reverse does not hold. Knowledge a priori 
is knowledge applicable to existence; but knowledge 
of the existent merely as such is not a priori. 

The significance of the theory of the a priori here 
presented may be brought out by considering the 
different ground of our knowledge in these two cases. 
We know that all parrots are birds without any ex- 
tended examination of parrots, or without any ex- 
perience of them at all, provided only we are clear 
about the concept which the term expresses. A par- ' 
rot which is not a bird is impossible because no mat- 
ter what characteristics any living creature might 
present, if it lacked the essential properties of a bird 
it would not properly be classifiable as a parrot. 
This a priori knowledge of parrots does not preclude 
the existence of any imaginable creature or limit in 
any way the possibilities of future experience. Liv- 
ing creatures not yet discovered, or not yet evolved, 
may be anything you please; but they cannot be 
parrots if they are not birds. It is for this reason 
and as is here maintained for this reason alone, 
that the proposition can be known true a priori. 
The illustration chosen is trivial ; but it will never- 
theless be typical provided only we remember that 
the distinction of real from unreal is a classifica- 
tion, and that what is designated as '^unreal" as well 
as the "real" is given in experience. Knowledge a 
priori is knowledge of our own concepts. It is also 
knowledge of reality in the sense that certain kinds 


of realities must exhibit certain categorial charac- 
teristics ; their failure to do so rules them out of the 
category in question. 

Since adequate examination of the distinction 
between propositions in intension and propositions 
in extension is nowhere to be found in the literature 
of logic, it may be of assistance to present here 
briefly the application of this distinction to the tra- 
ditional forms. 

The universal "All X is Y" in intension means, 
"If anything is an J5T, then necessarily it is a Y ; all 
possible JS?s are Y's." The universal in extension 
means, "All actual X*s are Y 9 s." It is particularly 
illuminating to consider the special case in which no 
X 9 s exist. Contrast "All trespassers on this prop- 
erty are liable to arrest" with "All trespassers on 
this property are minors," Suppose that nobody 
trespasses. The truth of the former is unaffected. 
If certain conditions designated by law have been 
met, trespass implies liability to arrest. The latter 
proposition, one might say, becomes insignificant. 
But every proposition ought to be true or false un- 
der all circumstances; and logicians have had to be 
precise about this matter because of certain prob- 
lems, some of which will be mentioned in what fol- 
lows. According to the dictum now commonly ac- 
cepted, if nobody trespasses, then the proposition, 
"All trespassers are Y" is true, whatever Y may be ; 
or in general, if the subject-term denotes an empty 
class, every universal proposition about that sub- 
ject is true in extension. (Many logicians neglect the 
distinction of extension and intension, and so apply 
this dictum to all universal propositions, with con- 
sequent difficulties to their theories.) If nobody tres- 
passes, then in extension any proposition whatever 
is true about all trespassers ; all the trespassers that 
exist are minors, red devils, pink elephants, or what 
you will. 

The case of the universal negative is somewhat 


more obvious. In intension, "No X is Y" means, 
"The concept X excludes the concept Y\ no pos- 
sible X is a Y." In extension, it means simply, "The 
class of things which are both X and Y is an empty 
class ; no such exist." Thus if nobody trespasses, 
"No trespasser is liable to arrest" would be true if 
meant in extension; but it is false in intension. In 
extension, .no trespasser is anything, because there 
are none. In intension, the question is whether the 
concept of trespass excludes liability to arrest, and 
that question is entirely independent of the exist- 
ence or non-existence of trespassers. 

Though the distinction of extensional from inten- 
sional meanings commonly passes unnoticed, it may 
be remarked for universal propositions in the case of 
definitions and legal principles. For particular 
propsitions, the distinction is even less familiar. But 
the ground of it is the same. In extension, "Some X 
is Y" means, "Members of the class *both X and Y 9 
exist." In intension it means the contradiction of the 
universal negative, "The concept X does not ex- 
clude the concept Y ; the logical species 'both X and 
Y 9 is possible or conceivable." Similarly, the par- 
ticular negative, "Some X is not Y 9 " means in ex- 
tension, "Members of the class 'X but not Y* exist'* ; 
while in intension it means, "The concept X does not 
imply the concept Y ; the logical species *X but not 
Y 9 is conceivable*" 

Though as has been said, particular propositions 
which are intensional in meaning are relatively in- 
frequent, still examples may be found. For instance, 
if a professor of jurisprudence should say, "Some 
grounds of civil suit are torts and some are not," 
and should be asked for demonstration, he might 
give it without any recourse to actual breaches of 
contract, conspiracies to defraud, etc., merely by 
pointing out the implications of legal concepts. His 
meaning is that, under the law, there are two species 
of possible action for which a remedy may be sought 


by civil suit. Or again, he might prove his point by 
purely hypothetical cases ; that they never occurred 
would be irrelevant. 

The case of the singular proposition is especially 
interesting. Mr. Russell has propounded the theory 
that a proposition about a singular subject (de- 
scribed by some phrase) is true if and only if (1) 
everything to which the subject-term applies has 
the predicated character, (2) the subject-term ap- 
plies to one existent thing, (3) it applies to only 
one. This has gained wide currency among logicians. 
It is a doubtful interpretation of the singular propo- 
sition in extension, and a perfectly impossible one 
of the singular in intension. 

The doubt as to its correctness in extension is un- 
important for us but may be noted in passing.* If 
a boastful friend asserts, "All the fish I catch will 
be big ones," we might rally him by rejoining, "All 
the fish you catch you can put in your eye." This 
answer, in the form of a universal proposition, is 
meant in extension. The point of making it is that 
we assert, by implication, that he will not catch any 
fish, because otherwise our proposition would be 
false. Thus we illustrate the fact that common-sense 
recognizes that the universal proposition in exten- 
sion is true when the subject denotes an empty class. 
But if our friend should say instead, "The first fish 
I catch will be a big one" (a singular proposition), 
we might similarly rejoin, "The first fish you catch 
will be a whale," meaning that this is true because 
there will not be any first fish. Thus if the inter- 
pretation of the singular proposition is to accord 
with some current usages, it must, like the univer- 
sal, be true whenever the subject denotes the non- 
existent, instead of false as Mr. Russell's theory re- 

The singular in intension, while infrequent, is of 

pretty clear interpretation. If I say, "The President 

*I mention only one ground of doubt; others have been put forward. 


of the United States must be native-born," my ob- 
vious meaning is to assert that the concept "Presi- 
dent," as delimited by the Constitution, has this im- 
plication. If it should happen that the incumbent of 
the office had died and no successor had been in- 
ducted, that would be irrelevant to my intended 
assertion. Similarly, a lawyer might state, "The 
residuary legatee under this will, is entitled to 
. . .," meaning to explicate the terms of the will 
under the law, and meaning nothing whatever about 
the existence or non-existence of a party answering 
to the designation of residuary legatee under the , 
will. The singular proposition in intension, like the 
universal, states that the subject-concept implies 
(or does not imply) the predicate. The difference is 
only that the subject of the singular in intension is 
a concept such that it applies to one, and to only one, 
possible object. 

Thus, throughout, the proposition in intension is 
true if it correctly states a relation of concepts, and 
is entirely independent of considerations of existence 
or non-existence. Propositions in extension, on the 
contrary, are dependent on facts of existence. Thus 
propositions in intension are analytically determin- 
able and a priori ; propositions in extension are syn- 
thetic and a posteriori. 

In conclusion, some of the logical difficulties which 
traditional logical precepts encounter in the case of 
extensional propositions the subject of which denotes 
an empty class, may be briefly noted. According to 
tradition, the relations of the four typical proposi- 
tions, (A) "All X is F," (J5) "No X is F, (I) 
"Some X is F," and (0) "Some X is not F," are as 
follows : A and E are contraries, that is, such that 
they cannot both be true but both may be false. I 
and O are subcontraries, that is, they may both be 
true but cannot both be false. A and 0, E and /, 
are contradictories, that is, they cannot both be 


true and cannot both be false. / follows from A, 
and O from E 9 but the reverse implications do not 
hold. Let "-5T is Y" be "trespassers will be prose- 
cuted," and suppose nobody trespasses. "Some tres- 
passers will be prosecuted" and "Some trespassers 
will not be prosecuted" are then both false. If it 
should be said that "All trespassers will be prose- 
cuted" is false because there are none, then A and O 
will not be contradictory, since O is false also. E is 
obviously true ; hence E and / may be contradictory, 
But O does not follow from E 9 since E is true but O 
is false. For uniformity of relation between univer- 
sals and particulars, it is necessary to accept the 
current dictum, that A> like E 9 is true when the sub- 
ject denotes an empty class. This makes A and O y E 
and /, contradictory. But A and E are not con- 
trary, being both true ; / and O are not subcontrary, 
both being false. And J does not follow from A 9 nor 
O from -E, because A and E are both true, / and O 
both false. 

The traditional relations of the "square of oppo- 
sition" hold of propositions in intension, that is, of 
propositions about the all-possible. The relation of 
contradiction holds in extension on the current in- 
terpretation of A as always true when the subject 
denotes an empty class. The other traditional rela- 
tions all fail concerning propositions in extension. 

Traditional logical doctrines will uniformly be 
found to have been worked out for intension, and to 
be commonly applied as if all propositions stated re- 
lations of intension. Current revisions of tradition 
will all too frequently be found to have been formu- 
lated with an opposite oversight as if all proposi- 
tions had their meaning in extension. 



(Where references under an entry are numerous, the most impor- 
it or decisive is sometimes indicated by "esp.") 

tant or 

Abstraction, 47 jf., 54 jf., 66, 71, 

80, 89, 91, 101, 115, 184, 249. 

358, App. A passim, and App. 

E passim. 
Acquaintance, knowledge by. 

118 jf., esp. 183, 188, 292. 
Activity, 34, 41, 46, 50, 117, 119, 

136, esp. 140 jf., 403, 407, esp. 

Analysis, 5, 30, 33 jf ., 55 /., 78, 

81,99, esp. 106 Jf. 
Analytic proposition, 37, 231, 

237, esp. 240 /., 245, 303 ff. 9 

312, App. F passim. 
Animal faith, 310, 340, 379, 381. 
Appearance, 9, 134, 137, 152, 

290 jf., 338, 336, 348, 354, 359 

/., 369 jf., 424 Jf. 
ARISTOTLE, 27, 41. 
Art (see also Esthetic), 53. 
Attention, 59. 
Awareness, 37, Ch. II passim, 

120, 124, 135, Ch. IX passim, 


Beautiful, see Esthetic. 
Behavior, 20, 78 f 85 ff ., Ch. IV 

passim, 117, 239, 409, 420. 
BERGSON, H., 39, 41, 58, 147 /., 

410, 431. 
BERKELEY, G., 28 /., 81, 39, 44, 

91, 133, 165, 170. 190. 
BERNAYS, N., 244. 
Body, 234 ff. 
BORN, M., 893. 

"BOSANQUET, B., 81, 198. 

BROAD, C, D., 15. 31, 61 jf. 

Cause. 100, 234 /., 888, 426. 
Certainty, Ch. X passim, 847 /., 

369, 377. 
Change, 137 jf.. 141, 396 jf. 


Circularity, 82, 209 /. 

Classification, 4, 12, 49, 134, 237, 
247, 256, 268, 321. 

Clearness, 3, 86 Jf. 

Color, 171 /., 363 Jf., 430. 

Common sense, esp. 2, 18 /. 

Connotation, see Meaning. 

Consistency, 18, 22, 34, 85, 208 
jf., 237, 245 jf. 

Context, 144 /., 405, 413; the- 
ory of meaning, 69 /. 

Cosmology, see Metaphysics. 

Critical realism, 60, 155 /. 

Datum, 29/., 33, 39, 45 Jf., esp. 

53 ff., 190, 217. 831 /., 337 /^ 

418 jf. 
Deduction, 23, 202, 258, 394; 

of the categories, 38, 320. 
Deductive system, 105 jf., 249. 

397 jf., App. E passim. 
Definition, 12/., 16, 37, 67, esp. 

79 /., 231, 247, 256, 263 /., 437. 
Denotation, 78, 84, 121, 131, 249, 

407, 429 /. 
DESCARTES, B., 1, 154 /., 166, 

205, 235. 

Dialectical method, 20^. 
Ding an sich, see Things in them- 

Dream, see Illusion. 
Duration (see also Time), 41, 74. 

Effect, 127, 133 /., 152, 418, 426. 
Einfvhlung, see Empathy. 
EINSTEIN, A*, 254 if., 393. 
Empathy, 41, 409, 411. 
Empirical, 34, 37, 223, 263 jf ., 

Chs. IX and X passim, 416, 

App. E passim* 
Empiricism, 27 jf., 432. 



End, 2, 145, 410. 

Enjoyment, 120, 135. 

Error, 39, 42J\, 121, 131 /., esp. 

157-165, 176 jf. 

Essence, 60, 62, 72, 120 /., 130. 
Essential (see also Definition), 

374 /. 
Esthesis, 53 /., 75, 275 /., App. 

B passim t 407. 
Esthetic 27, 112, 118, 120, 135, 

147, App. B passim, 407. 
Esthetics, 29, App. B passim, 409. 
Ethics, 3, 14, 17, 112, 409 ff. 
Event, 59 /., 152 /., 307. 
Experience, esp, 59 ff. 
Explicative, see Analytic. 
Extension (see also Denotation) ; 

proposition in , App. F 


Falsity, see Truth and Error. 
FICHTE, J. G., 46, 190, 236, 425. 
Function, 170, 175, 186 ff. 

Generalization, 29, 134, 137 ff. ., 
292, Ch. X passim, 346 jf., 361, 
370 jf., 422, 434. 

Good, 2, 36, 408. 

GOBGIAS, 881. 

GREEN, T. H., 45, 

Hallucination, see Illusion. 
HAMILTON, W. R., 252. 
History, 67/., 152, 228, 267 ff. t 


HOLT, E. B., 43. 
HUME, D., 1, 28, 44, 195, 213, 

312, 319 Jf., 381. 
Hypothesis, 131 /., 142, 256 jf., 

3S6> 357, 4Z6jf. 

Idea, 9, 96, App. C passim. 
Idealism, 39, 45 Jf., 91, 149, 155, 

167, esp. 180 Jf., 410 /., 414. 
Identity, 125 /., 137, 411. 
Ignorance, 176 jf. 
Illusion (see also Error), 28, 57, 

138, 143, 225, 261, 299. 350 /. 
Image, 78, 81, 177. 
Imagination, 26, 162. 
Immediacy, 39 /., 44, 46, Ch* II 

passim, 91, 120, 147, 275 /., 

338, 358, App. B passim, 407, 


Inconsistency, see Consistency. 
Independent object, see under 


Induction, 23, 204, 323, 345, 400. 
Ineffable, 53, 112, 123, 126 /., 

144, 147 /. 

Intelligence, 110, 113 jf. 
Intelligibility, 144^., 222/., 348, 

Ch. XI passim, 396. 
Intention (see also Meaning); 

propositions in , App. F 

Interest, 51, 91, 110, 112 ff., 128, 

159, 265, 374, 407, 413. 
Interpretation, 5, 12, 14, 18, 27, 

30, esp. 48 jf., 58, 60, 268 Jf. 
Introspection, 73. 
Intuition, 41, 47, 66, 75 jf., Ill, 

121, 147 /., 198, 231, 241 Jf., 

297, 320, 404. 

JAMES, W., 65. 

Judgment (see also Prediction), 
46, 157-165, 245, Chs. IX and 
X passim. 408. 

KANT, I., 1, 8, 9, 55, 139, 149, 
151, 154, 214 /., 241, 295 jf. 

Language, 31, 33, 84, 87, 102, 

110.jf., 117 /., 123. 
LAPLACE, P. S., 32. 
Law (see also Generalization)* 

130, 254, 258, 262 jf., 321, 332 

jf.,S43, 373 jf. 
LEIBNITZ, G. W. v., 113. 
Like-minded, 21, 113 /. 
Limitations, 74 jf., 158, 176 if., 

319, 368, 373 J. 
LOCKE, J., 1, 31. 


Logic, 8, 14, 17, 69, 71, esp. 207 
jf., 245 jf., 306 /., App. F 

Mathematics, 95, 99, 104 /., 107, 
203, esp. 243 jf., App. E passim. 

Meaning, 40 jf., Chs. HI and IV 
passim, esp. 67-72, 197, 268 /., 
287, 408; of presentations, 
135 / 144. 



Memory, 51, 53, 125, 141. 
Metaphysics, Ch. I passim, 65, 


MILL, J.S., 250 Jf. 
Mistake, see Error. 
MONTAGUE, W. P., 43. 
Moral (see also Ethics), 3, 147. 
Mysticism, 40 jf., 47, 146, 406. 

Name, 121 jf., 132, 134 jf., 278 

jf., 294 /., 318. 
Natural light, 25, 198. 
Necessary, 195 /., Ch. VD1 pas- 

sim, 293, 324. 
Need, see Interest. 
New realism, 89, 42, 71 Jf., 135. 

Order, 107 /., 130, 143, 254, 341, 
Ch. XI passim, 397. 

Paradox, 149 jf., 424. 
Particular, 223, 283 jf., 301, 315 

jf., 336, 388. 
Passive consciousness, 30, 41 1 

137, esp. 141 Jf., 214 /. 
PEIBCE, C. S., 65, 133, 417 /. 
Perception, 42/., 73 jf., 134, 414. 
Permanence (see also Time), 33, 


PERBT, R. B., 149. 
Personality, 229, 410. 
Phenomenalism, 154 /., 165, 

Physical, 27, 161, 235, 252, 305, 

322, 393. 

PLATO, 21, 45, 70, 241, 309, 315. 
Postulate, 105, 243 jf., 409 ff. 
Practical, 18, 21, 31, 85, 119, 

123, 147, 343, 405. 
Pragmatic, 31, 34 /., 145 /., Ch. 

VTII passim, 300, 375, 410, 

Predication, Ch. IX passim, esp. 

278 ff, 
Prediction, 37, 44, Ch. V passim, 

277 Jf., 335, 375, 382 jf., 407. 
Present (see also Time); specious 

, 58, 61, 141. 
Presentation, 44, 46, 49, 58, esp. 

59 /., 63, 120, 134 /., 158, 278 

ff., 292, 310, 372, 413 /. 
Presupposition, 1, 198 ff., 415 Jf. 

Principia Mathematica, 95, 101, 

107, 244, 248. 
Property, 121 Jf., 130, 171, 234 

f., 283 Jf., 373 Jf., 396. 
Psychology, 6, 35, 55 Jf., 62, 68 

jf., 87, 160, 420. 
Purpose, 51, 93, 111, 119, 228, 

265 /., 272, 420. 

Quale, esp. 60 ff. and 121 /., 310, 
383, 887 Jf. 

Ratio cognoscendi, 149, 425 /. 

Rational, 22, 341 /., 347, 381 /. 

Rationalism, 24 ff., 33, 92, 202 
ff., 211 /., 233, 237, 432. 

Real, 8, 10, 16, 27, 36, 279, 413, 

Realism (see also New realism. 
Critical realism, and Represen- 
tationalism), 154 /., 166, 414. 

Reality, Ch. I passim, 112> 321, 
358, 366, 381, 390; indepen- 
dent . , 47, 64, 192 ff., 410. 

Reason (see also Rational), 20, 
92 jf., 99, 101, 233, 237. 

Receptivity (see also Intuition), 
42, 47, 217 Jf. 

Recognition, 134 Jf., 291, 294, 

Reflective method, 22 Jf., 56. 

Regulative principle, 353, 366. 

Relations, 40, 82, 91, 98. 

Religion, 112, 147, 409 Jf. 

Representationalism, 43, 135, 
Ch. VI passim. 

Right (see also Ethics), , 27, 36. 

ROYCE, J., 86, 101. 

RUSSELL, B. A. W., 15, 64, 95, 
181, 206, 244, 438. 


Science, 2, 6, 32, 56, 70, 99, 147, 
235 /., 249, 254, 311, 342, 374 

J., App. A passim. 
conscious, 10, 18, 88. 
Self-contradictory, 205 Jf. 
Sensum, 81, 61 ff. 
Shape, 172 /. 
SHEFFER, H. M., 244. 
Sign (cognitive), 44, 119 ff., 144, 

192, 402, 405 /. 
Similarity, 92, 364 /. 



Size, 168 jf. 

Skepticism, 24, 44, 72 /., 116, 

155 /., 195, 213, 215 ff n Miff., 

311 /., 819 jf., Ch. XI passim. 
Social, 21, 25, Ch. IV passim, 

117, 147, 228, 268 /.. 408 /. 
Solipsism, 183, 410. 
Sophists, 156. 
Space, 82/., 140, 148, 234 /., 241 

/, 297 fa SIS ff.i Non- 
Euclidean , 253 JT. 
Speculation, 6/ 
Speech, see Language. 
Stream of consciousness, 58 /., 

141, 357. 

Subject (of knowledge), 47, 164. 
Subjective, 112, 116, 121, 147, 

329, 348, 410. 
Substance, 5, 100, 395 /. 
Substantive, 67, 70, 99, 106, 369, 

374, 397. 
Synthetic proposition, 245, 303 

ff., 312, App. F passim; a 

priori, 1, 245. 

Thing, 58, 136 ff.. 234 /., 283 /., 
307, 320, 348, 366 ff., 372 jf ., 
App. A passim. 

Things in themselves, 64, 156, 
179 jf., 214JF..415. 

Time, 51, 60/.. 130, 140 /,. esp. 

148 #, 234 /., 255 #, 281. 
Transcendent, 20, 25, 30, 34 /., 

40, 45, 71, 93, 176, 183, 223, 

228, 231 Jf., 412 /., 418, 425. 
Truth, 25, 39, 157-165, 206 /., 

231, 240 / esp. 267 ff., 330 

Jf. 9 341, 346 /., 382. 

Uniformity, 137, 257, 261, Ch. 
XI passim, App. A passim. 

Universal; concepts, 61, 121 
Jf.; propositions, 199, 231, 
315 ff. f 336, 436 J.; abstract 
, 10; concrete , 9, 55. 

Validity (see also Judgment), 2, 

11, 36, 325, 331, 338 ff., 346, 

376, 886. 
Value, 18, 51, 71, 89, 145, 248, 

App. B passim, 407. 
Veridical, 28, 42/., 282. 
Verification, 124 Jf., 130 JF., 149 

jf ., 58. 

WHITEHEAD, A, N., 15, 95, 152, 

244, 431. 
Will, see Purpose. 

ZENO, 148.